{ "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "bdu-stMDC9lL", "outputId": "358497ce-924b-49d1-9fc0-9008467c2dfb" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Collecting sodapy\n", " Downloading sodapy-2.2.0-py2.py3-none-any.whl (15 kB)\n", "Requirement already satisfied: requests>=2.28.1 in /usr/local/lib/python3.10/dist-packages (from sodapy) (2.31.0)\n", "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests>=2.28.1->sodapy) (3.3.2)\n", "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests>=2.28.1->sodapy) (3.6)\n", "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests>=2.28.1->sodapy) (2.0.7)\n", "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests>=2.28.1->sodapy) (2024.2.2)\n", "Installing collected packages: sodapy\n", "Successfully installed sodapy-2.2.0\n" ] } ], "source": [ "# !pip install requests\n", "!pip install sodapy\n", "\n", "import requests\n", "from requests.auth import HTTPBasicAuth\n", "\n", "## Understanding, Exploring and Visualizing Data\n", "import pandas as pd\n", "import seaborn as sns\n", "import matplotlib\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "from sklearn import datasets\n", "from sodapy import Socrata\n", "\n", "\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.ensemble import RandomForestRegressor\n", "from sklearn.preprocessing import LabelEncoder\n", "from sklearn.metrics import mean_squared_error\n", "\n", "\n", "import pandas as pd\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import StandardScaler\n", "from sklearn.preprocessing import MinMaxScaler\n", "from sklearn.tree import DecisionTreeClassifier\n", "from sklearn.ensemble import GradientBoostingClassifier\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.svm import LinearSVC\n", "from sklearn.neighbors import KNeighborsClassifier\n", "from sklearn.metrics import accuracy_score, confusion_matrix\n", "import pandas as pd\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.metrics import mean_squared_error\n", "from sklearn.preprocessing import OneHotEncoder\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.pipeline import Pipeline\n", "import dask.dataframe as dd\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "cry5oVSCGK-X", "outputId": "76cb8110-2457-4838-f590-976f79259033" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "\n", "RangeIndex: 200000 entries, 0 to 199999\n", "Data columns (total 45 columns):\n", " # Column Non-Null Count Dtype \n", "--- ------ -------------- ----- \n", " 0 bble 200000 non-null object\n", " 1 boro 200000 non-null object\n", " 2 block 200000 non-null object\n", " 3 lot 200000 non-null object\n", " 4 owner 191502 non-null object\n", " 5 bldgcl 200000 non-null object\n", " 6 taxclass 200000 non-null object\n", " 7 ltfront 200000 non-null object\n", " 8 ltdepth 200000 non-null object\n", " 9 stories 163223 non-null object\n", " 10 fullval 200000 non-null object\n", " 11 avland 200000 non-null object\n", " 12 avtot 200000 non-null object\n", " 13 exland 200000 non-null object\n", " 14 extot 200000 non-null object\n", " 15 excd1 56136 non-null object\n", " 16 staddr 199329 non-null object\n", " 17 bldfront 200000 non-null object\n", " 18 blddepth 200000 non-null object\n", " 19 avland2 144671 non-null object\n", " 20 avtot2 144671 non-null object\n", " 21 exland2 30762 non-null object\n", " 22 extot2 46378 non-null object\n", " 23 period 200000 non-null object\n", " 24 year 200000 non-null object\n", " 25 valtype 200000 non-null object\n", " 26 zip 175469 non-null object\n", " 27 exmptcl 4803 non-null object\n", " 28 easement 2872 non-null object\n", " 29 borough 166533 non-null object\n", " 30 latitude 166417 non-null object\n", " 31 longitude 166417 non-null object\n", " 32 community_board 166417 non-null object\n", " 33 council_district 166417 non-null object\n", " 34 census_tract 166417 non-null object\n", " 35 bin 165961 non-null object\n", " 36 nta 166417 non-null object\n", " 37 geocoded_column 166417 non-null object\n", " 38 :@computed_region_efsh_h5xi 165123 non-null object\n", " 39 :@computed_region_f5dn_yrer 166404 non-null object\n", " 40 :@computed_region_yeji_bk3q 166404 non-null object\n", " 41 :@computed_region_92fq_4b7q 166404 non-null object\n", " 42 :@computed_region_sbqj_enih 166404 non-null object\n", " 43 ext 11668 non-null object\n", " 44 excd2 3508 non-null object\n", "dtypes: object(45)\n", "memory usage: 68.7+ MB\n" ] } ], "source": [ "# First Dataset (2010 - 2019)\n", "# https://data.cityofnewyork.us/City-Government/Property-Valuation-and-Assessment-Data/yjxr-fw8i/about_data\n", "api_endpoint = \"https://data.cityofnewyork.us/resource/yjxr-fw8i.json\"\n", "api_key = \"7cdpwsl5p8eqeg1d24xsbq64w\"\n", "api_secret = \"31grq5eswi42uh3k9hu6eyjlkp3x0wb0bsyk4r9yu2jz2mr6b9\"\n", "\n", "response = requests.get(api_endpoint, auth=HTTPBasicAuth(api_key, api_secret))\n", "\n", "limit = 1000 # Number of items per request\n", "offset = 0 # Starting point for each request\n", "# Total records in the file is 9845857, we are omitting 2 records\n", "total_records = 9845855 # Total records in the dataset (adjust as needed)\n", "\n", "# Initialize an empty DataFrame to store all data\n", "all_data = pd.DataFrame()\n", "i = 0\n", "\n", "# We are only taking the first 60k data\n", "# 20022\n", "2\n", "while i < 200:\n", " i += 1\n", " # Use the API call to request the server for info\n", " response = requests.get(api_endpoint, auth=HTTPBasicAuth(api_key, api_secret),\n", " params={\"$limit\": limit, \"$offset\": offset})\n", "\n", " # Check if the request was successful\n", " if response.status_code == 200:\n", " # Load current data into a DataFrame\n", " current_data = pd.DataFrame(response.json())\n", "\n", " # Concatenate the current data to the main DataFrame\n", " all_data = pd.concat([all_data, current_data], ignore_index=True)\n", "\n", " # Update the offset for the next request\n", " offset += limit\n", " #print(current_data)\n", " else:\n", " print(f\"Failed to fetch data: {response.status_code}\")\n", " print(response.text)\n", " break\n", "\n", "# Save the DataFrame to a CSV file\n", "all_data.to_csv('nyc_data1.csv', index=False)\n", "\n", "# Display the shape of the DataFrame to verify the number of rows loaded\n", "# print(all_data.shape)\n", "\n", "\n", "df1 = all_data\n", "df1.info()" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 236 }, "id": "no0qDem4GK4A", "outputId": "2b748f85-2b0b-4e7a-ded5-b8fcfbcf7e0e" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " parid boro block lot rectype year secvol pymktland pymkttot \\\n", "0 3025410030 3 2541 30 1 2021 901 87000 1721000 \n", "1 3025410033 3 2541 33 1 2021 901 106000 2562400 \n", "2 3025410036 3 2541 36 1 2021 901 87000 1939200 \n", "3 3025410038 3 2541 38 1 2021 901 314000 1593000 \n", "4 3025410039 3 2541 39 1 2021 901 298000 1942000 \n", "\n", " pyactland ... cbntaxclass fintaxclass easement reuc_ref subident_reuc \\\n", "0 39150 ... NaN NaN NaN NaN NaN \n", "1 47700 ... NaN NaN NaN NaN NaN \n", "2 39150 ... NaN NaN NaN NaN NaN \n", "3 18840 ... NaN NaN NaN NaN NaN \n", "4 17880 ... NaN NaN NaN NaN NaN \n", "\n", " ident subident roll_section reuc_description valref \n", "0 NaN NaN NaN NaN NaN \n", "1 NaN NaN NaN NaN NaN \n", "2 NaN NaN NaN NaN NaN \n", "3 NaN NaN NaN NaN NaN \n", "4 NaN NaN NaN NaN NaN \n", "\n", "[5 rows x 132 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
paridboroblocklotrectypeyearsecvolpymktlandpymkttotpyactland...cbntaxclassfintaxclasseasementreuc_refsubident_reucidentsubidentroll_sectionreuc_descriptionvalref
0302541003032541301202190187000172100039150...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
13025410033325413312021901106000256240047700...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
2302541003632541361202190187000193920039150...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
33025410038325413812021901314000159300018840...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
43025410039325413912021901298000194200017880...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
\n", "

5 rows × 132 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df2" } }, "metadata": {}, "execution_count": 63 } ], "source": [ "# Second Dataset (NEW) (2021-2023)\n", "# https://data.cityofnewyork.us/City-Government/Property-Valuation-and-Assessment-Data-Tax-Classes/8y4t-faws/data_preview\n", "\n", "api_endpoint = \"https://data.cityofnewyork.us/resource/8y4t-faws.json\"\n", "api_key = \"6iicdls0lllux5k2zh4wmcbcy\"\n", "api_secret = \"nol8pwnk4foppzbxrm8vl23sgljt0mqzkxpzo0xscgbxs1zv9\"\n", "\n", "response = requests.get(api_endpoint, auth=HTTPBasicAuth(api_key, api_secret))\n", "\n", "limit = 1000 # Number of items per request\n", "offset = 0 # Starting point for each request\n", "# Total records in the file is 9845857, we are omitting 2 records\n", "total_records = 9845855 # Total records in the dataset (adjust as needed)\n", "\n", "# Initialize an empty DataFrame to store all data\n", "all_data = pd.DataFrame()\n", "i = 0\n", "\n", "# We are only taking the first 60k data\n", "# 200\n", "while i < 200:\n", " i += 1\n", " # Use the API call to request the server for info\n", " response = requests.get(api_endpoint, auth=HTTPBasicAuth(api_key, api_secret),\n", " params={\"$limit\": limit, \"$offset\": offset})\n", "\n", " # Check if the request was successful\n", " if response.status_code == 200:\n", " # Load current data into a DataFrame\n", " current_data = pd.DataFrame(response.json())\n", "\n", " # Concatenate the current data to the main DataFrame\n", " all_data = pd.concat([all_data, current_data], ignore_index=True)\n", "\n", " # Update the offset for the next request\n", " offset += limit\n", " #print(current_data)\n", " else:\n", " print(f\"Failed to fetch data: {response.status_code}\")\n", " print(response.text)\n", " break\n", "\n", "# Save the DataFrame to a CSV file\n", "all_data.to_csv('nyc_data2.csv', index=False)\n", "\n", "# Display the shape of the DataFrame to verify the number of rows loaded\n", "# print(all_data.shape)\n", "\n", "\n", "df2 = all_data\n", "df2.head()" ] }, { "cell_type": "markdown", "metadata": { "id": "AJwJdGMrh3tu" }, "source": [ "\n", "Dataset common:\n", "\n", "Dataset1:Dataset2\n", "\n", "2010 : 2025\n", "\n", "\n", "\n", "---\n", "\n", "\n", "boro : boro (borough)\n", "\n", "block : block (block range by borough)\n", "\n", "lot : lot (Unique number within BORO/BLOCK)\n", "\n", "easement : easement (codes for things such as A for air rights and R for railroads)\n", "\n", "year : year\n", "\n", "FINTAXCLASS : taxclass\n", "\n", "bldg_class : bldgcl (building class)\n", "\n", "owner : owner\n", "\n", "street_name : staddr (street name : street address (is there a diff))\n", "\n", "LOT_FRT : LTFRONT (lot frontage : lot width)\n", "\n", "LOT_DEP : ltdep (lot depth)\n", "\n", "bld_story : stories (number of stories)\n", "\n", "\n", "PYMKTLAND : FULLVAL Market Assessed Land Value\n", "\n", "PYACTLAND : AVLAND Actual Assessed Land Value\n", "\n", "PYACTTOT : AVTOT Actual Assessed Total Value\n", "\n", "\n", "\n", "\n", "---\n", "\n", "\n", "Dataset 1 Useful:\n", "\n", "\n", "(These values are originally assessed)\n", "\n", "PYMKTLAND Market Assessed Land Value\n", "\n", "PYMKTTOT Market Assessed Total Value\n", "\n", "PYACTLAND Actual Assessed Land Value\n", "\n", "PYACTTOT Actual Assessed Total Value\n", "\n", "PYTAXCLASS Property Tax Class\n", "\n", "\n", "(These values are finally assessed)\n", "\n", "FINMKTLAND Final Market Assessed Land Value\n", "\n", "FINMKTTOT Final Market Assessed Total Value\n", "\n", "FINACTLAND Final Actual Assessed Land Value\n", "\n", "FINACTTOT Final Actual Assessed Total Value\n", "\n", "FINTAXCLASS Property Tax Class\n", "\n", "\n", "(These values are currently assessed)\n", "\n", "CURMKTLAND Current Market Assessed Land Value\n", "\n", "CURMKTTOT Current Market Assessed Total Value\n", "\n", "CURACTLAND Current Actual Assessed Land Value\n", "\n", "CURACTTOT Current Actual Assessed Total Value\n", "\n", "\n", "noav A building in progress\n", "\n", "stcode street code (idk)\n", "\n", "\n", "\n", "\n", "---\n", "\n", "\n", "\n", "Dataset 2 Useful:\n", "\n", "FULLVAL Market Value\n", "\n", "AVLAND Actual Land Value\n", "\n", "AVTOT Actual Total Value\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Py1S2O6-f1VA", "outputId": "ebc39960-585f-414a-e18b-4e9d7432b246" }, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ ":5: DtypeWarning: Columns (0,62,66,69,115,120,122,123,124,125,126,127,128,130,131) have mixed types. Specify dtype option on import or set low_memory=False.\n", " df2 = pd.read_csv('nyc_data2.csv')\n" ] } ], "source": [ "# Read from CSV\n", "\n", "df1 = pd.read_csv('nyc_data1.csv')\n", "\n", "df2 = pd.read_csv('nyc_data2.csv')\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "LI7VStz0tRXJ", "outputId": "0abbb4d6-c8fa-42bf-954a-8d90a393c1ac" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "0 87000\n", "1 106000\n", "2 87000\n", "3 314000\n", "4 298000\n", " ... \n", "199995 212000\n", "199996 168000\n", "199997 148000\n", "199998 188000\n", "199999 238000\n", "Name: pymktland, Length: 200000, dtype: int64" ] }, "metadata": {}, "execution_count": 65 } ], "source": [ "df2['pymktland']" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "PbSEQazKhS0I", "colab": { "base_uri": "https://localhost:8080/" }, "outputId": "f4d9ace4-4edc-4bd8-d707-f278b2fdfbb3" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ " boro block lot easement year taxclass bldgcl owner \\\n", "0 3 2541 30 NaN 2021 NaN C1 299 LLC \n", "1 3 2541 33 NaN 2021 NaN C1 293 LLC \n", "2 3 2541 36 NaN 2021 NaN C1 287A LLC \n", "3 3 2541 38 NaN 2021 NaN B9 JUSZCZAK, ANTONINA \n", "4 3 2541 39 NaN 2021 NaN B9 SAWICKI CZESLAW \n", "... ... ... ... ... ... ... ... ... \n", "199284 3 7820 15 NaN 2021 NaN B1 VELEZ, LUIS A \n", "199285 3 7820 17 NaN 2021 NaN B1 GEDEON GINETTE \n", "199286 3 7820 18 NaN 2021 NaN B1 MARK, GERARD \n", "199287 3 7820 19 NaN 2021 NaN B1 BAHL PHILIP \n", "199288 3 7820 21 NaN 2021 NaN B3 BENNETT, ROBERT \n", "\n", " staddr ltfront ltdepth stories fullval avland \\\n", "0 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "1 MCGUINNESS BOULEVARD 76.00 35.00 5.0 106000 47700 \n", "2 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "3 JAVA STREET 25.00 100.00 2.0 314000 18840 \n", "4 JAVA STREET 25.00 100.00 3.0 298000 17880 \n", "... ... ... ... ... ... ... \n", "199284 KINGS HIGHWAY 35.00 114.00 2.0 212000 12720 \n", "199285 KINGS HIGHWAY 20.00 107.00 2.0 168000 10080 \n", "199286 KINGS HIGHWAY 20.00 94.33 2.0 148000 8880 \n", "199287 KINGS HIGHWAY 23.33 81.67 2.0 188000 11280 \n", "199288 KINGS HIGHWAY 91.00 52.33 2.0 238000 14280 \n", "\n", " avtot \n", "0 774450 \n", "1 1153080 \n", "2 872640 \n", "3 95580 \n", "4 116520 \n", "... ... \n", "199284 42120 \n", "199285 36540 \n", "199286 36540 \n", "199287 41880 \n", "199288 48420 \n", "\n", "[199289 rows x 15 columns]\n", " boro block lot easement year taxclass bldgcl owner \\\n", "0 3 2541 30 NaN 2021 NaN C1 299 LLC \n", "1 3 2541 33 NaN 2021 NaN C1 293 LLC \n", "2 3 2541 36 NaN 2021 NaN C1 287A LLC \n", "3 3 2541 38 NaN 2021 NaN B9 JUSZCZAK, ANTONINA \n", "4 3 2541 39 NaN 2021 NaN B9 SAWICKI CZESLAW \n", "... ... ... ... ... ... ... ... ... \n", "199284 3 7820 15 NaN 2021 NaN B1 VELEZ, LUIS A \n", "199285 3 7820 17 NaN 2021 NaN B1 GEDEON GINETTE \n", "199286 3 7820 18 NaN 2021 NaN B1 MARK, GERARD \n", "199287 3 7820 19 NaN 2021 NaN B1 BAHL PHILIP \n", "199288 3 7820 21 NaN 2021 NaN B3 BENNETT, ROBERT \n", "\n", " staddr ltfront ltdepth stories fullval avland \\\n", "0 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "1 MCGUINNESS BOULEVARD 76.00 35.00 5.0 106000 47700 \n", "2 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "3 JAVA STREET 25.00 100.00 2.0 314000 18840 \n", "4 JAVA STREET 25.00 100.00 3.0 298000 17880 \n", "... ... ... ... ... ... ... \n", "199284 KINGS HIGHWAY 35.00 114.00 2.0 212000 12720 \n", "199285 KINGS HIGHWAY 20.00 107.00 2.0 168000 10080 \n", "199286 KINGS HIGHWAY 20.00 94.33 2.0 148000 8880 \n", "199287 KINGS HIGHWAY 23.33 81.67 2.0 188000 11280 \n", "199288 KINGS HIGHWAY 91.00 52.33 2.0 238000 14280 \n", "\n", " avtot \n", "0 774450 \n", "1 1153080 \n", "2 872640 \n", "3 95580 \n", "4 116520 \n", "... ... \n", "199284 42120 \n", "199285 36540 \n", "199286 36540 \n", "199287 41880 \n", "199288 48420 \n", "\n", "[199289 rows x 15 columns]\n" ] } ], "source": [ "# This is to combine the values we want from df1 (2010-2019) and df2 (2021-2025), we have also cut the data to 2021-2023\n", "\n", "# This is because d1 has the years formatted like this 2010/11 when we want it as a signular year 2010\n", "df1['year'] = df1['year'].str.slice(0, 4)\n", "\n", "# Drop rows where the year is 2024 or 2025\n", "df2 = df2[~df2['year'].isin([2024, 2025])]\n", "\n", "\n", "columns_mapping = {\n", " 'boro': 'boro', # d2:boro -> d3:boro\n", " 'block': 'block', # d2:block -> d3:block\n", " 'lot': 'lot', # d2:lot -> d3:lot\n", " 'easement': 'easement', # d2:easement -> d3:easement\n", " 'year': 'year', # d2:year -> d3:year\n", " 'fintaxclass': 'taxclass', # d2:FINTAXCLASS -> d3:taxclass\n", " 'bldg_class': 'bldgcl', # d2:bldg_class -> d3:bldgcl\n", " 'owner': 'owner', # d2:owner -> d3:owner\n", " 'street_name': 'staddr', # d2:street_name -> d3:staddr\n", " 'lot_frt': 'ltfront', # d2:LOT_FRT -> d3:LTFRONT\n", " 'lot_dep': 'ltdepth', # d2:LOT_DEP -> d3:ltdep\n", " 'bld_story': 'stories', # d2:bld_story -> d3:stories\n", " 'pymktland': 'fullval', # d2:PYMKTLAND -> d3:FULLVAL\n", " 'pyactland': 'avland', # d2:PYACTLAND -> d3:AVLAND\n", " 'pyacttot': 'avtot' # d2:PYACTTOT -> d3:AVTOT\n", "}\n", "\n", "\n", "df3_columns = ['boro', 'block', 'lot', 'easement', 'year', 'taxclass', 'bldgcl',\n", " 'owner', 'staddr', 'ltfront', 'ltdepth', 'stories', 'fullval', 'avland', 'avtot']\n", "\n", "df3 = pd.DataFrame(columns=df3_columns)\n", "\n", "# Select columns safely from df1_renamed and df2\n", "# df2_renamed = df2.rename(columns={v: k for k, v in columns_mapping.items()})\n", "\n", "# Create a new dictionary with keys and values swapped\n", "'''\n", "swapped_columns_mapping = {}\n", "for original_column, new_column in columns_mapping.items():\n", " swapped_columns_mapping[new_column] = original_column\n", "\n", "# Use the new dictionary to rename columns in df2\n", "df2_renamed = df2.rename(columns=swapped_columns_mapping)\n", "print(df2_renamed)\n", "df2_selected = df2_renamed.reindex(columns=df3_columns)\n", "\n", "\n", "# df1_selected = df1.reindex(columns=df3_columns)\n", "\n", "'''\n", "# Step 1: Filter columns in df2 that are keys in the columns_mapping\n", "df2_filtered = df2[list(columns_mapping.keys())]\n", "\n", "# Step 2: Rename columns in df2 according to the mapping\n", "df2_renamed = df2_filtered.rename(columns=columns_mapping)\n", "\n", "# Step 3: Merge or Update\n", "# Option 1: Merge\n", "df3_merged = pd.merge(df3, df2_renamed, on=list(columns_mapping.values()), how='outer')\n", "print(df3_merged)\n", "# Option 2: Update\n", "df3.update(df2_renamed)\n", "print(df3_merged)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "L7uSGb6GbIXc", "outputId": "197d7fab-8daf-4599-c72b-0f76b9ffdc68" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl owner \\\n", "0 3 2541 30 NaN 2021 NaN C1 299 LLC \n", "1 3 2541 33 NaN 2021 NaN C1 293 LLC \n", "2 3 2541 36 NaN 2021 NaN C1 287A LLC \n", "3 3 2541 38 NaN 2021 NaN B9 JUSZCZAK, ANTONINA \n", "4 3 2541 39 NaN 2021 NaN B9 SAWICKI CZESLAW \n", "... ... ... ... ... ... ... ... ... \n", "199995 3 7820 15 NaN 2021 NaN B1 VELEZ, LUIS A \n", "199996 3 7820 17 NaN 2021 NaN B1 GEDEON GINETTE \n", "199997 3 7820 18 NaN 2021 NaN B1 MARK, GERARD \n", "199998 3 7820 19 NaN 2021 NaN B1 BAHL PHILIP \n", "199999 3 7820 21 NaN 2021 NaN B3 BENNETT, ROBERT \n", "\n", " staddr ltfront ltdepth stories fullval avland \\\n", "0 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "1 MCGUINNESS BOULEVARD 76.00 35.00 5.0 106000 47700 \n", "2 MCGUINNESS BOULEVARD 62.00 35.00 5.0 87000 39150 \n", "3 JAVA STREET 25.00 100.00 2.0 314000 18840 \n", "4 JAVA STREET 25.00 100.00 3.0 298000 17880 \n", "... ... ... ... ... ... ... \n", "199995 KINGS HIGHWAY 35.00 114.00 2.0 212000 12720 \n", "199996 KINGS HIGHWAY 20.00 107.00 2.0 168000 10080 \n", "199997 KINGS HIGHWAY 20.00 94.33 2.0 148000 8880 \n", "199998 KINGS HIGHWAY 23.33 81.67 2.0 188000 11280 \n", "199999 KINGS HIGHWAY 91.00 52.33 2.0 238000 14280 \n", "\n", " avtot \n", "0 774450 \n", "1 1153080 \n", "2 872640 \n", "3 95580 \n", "4 116520 \n", "... ... \n", "199995 42120 \n", "199996 36540 \n", "199997 36540 \n", "199998 41880 \n", "199999 48420 \n", "\n", "[199289 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.0035.005.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.0035.005.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.0035.005.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.00100.002.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.00100.003.029800017880116520
................................................
1999953782015NaN2021NaNB1VELEZ, LUIS AKINGS HIGHWAY35.00114.002.02120001272042120
1999963782017NaN2021NaNB1GEDEON GINETTEKINGS HIGHWAY20.00107.002.01680001008036540
1999973782018NaN2021NaNB1MARK, GERARDKINGS HIGHWAY20.0094.332.0148000888036540
1999983782019NaN2021NaNB1BAHL PHILIPKINGS HIGHWAY23.3381.672.01880001128041880
1999993782021NaN2021NaNB3BENNETT, ROBERTKINGS HIGHWAY91.0052.332.02380001428048420
\n", "

199289 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df2_renamed" } }, "metadata": {}, "execution_count": 67 } ], "source": [ "df2_renamed" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "OrO1dxjzYfsO" }, "outputs": [], "source": [ "# Delete any columns that arent defined in df3_columns\n", "df1 = df1[df3_columns]\n", "\n", "# Concatenate df1 and df2\n", "df3 = pd.concat([df2_renamed, df1], ignore_index=True)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "I6KgWYjCkUe4", "outputId": "d1f7e3df-c50c-4e62-de3d-cf655ee77860" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "0 3 2541 30 NaN 2021 NaN C1 \n", "1 3 2541 33 NaN 2021 NaN C1 \n", "2 3 2541 36 NaN 2021 NaN C1 \n", "3 3 2541 38 NaN 2021 NaN B9 \n", "4 3 2541 39 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "399284 2 3432 1941 NaN 2018 1A R3 \n", "399285 2 3332 42 NaN 2018 2 C7 \n", "399286 2 3385 38 NaN 2018 1 B1 \n", "399287 2 2963 6 NaN 2018 2B C1 \n", "399288 2 2928 33 NaN 2018 4 Z9 \n", "\n", " owner staddr ltfront ltdepth \\\n", "0 299 LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "1 293 LLC MCGUINNESS BOULEVARD 76.0 35.0 \n", "2 287A LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "3 JUSZCZAK, ANTONINA JAVA STREET 25.0 100.0 \n", "4 SAWICKI CZESLAW JAVA STREET 25.0 100.0 \n", "... ... ... ... ... \n", "399284 ALEMAN, CYNTHIA 116 SURF DRIVE 0.0 0.0 \n", "399285 3105 DECATUR ASSOCIAT 3105 DECATUR AVENUE 125.0 119.0 \n", "399286 HOGAN, RUTH 345 EAST 236 STREET 25.0 100.0 \n", "399287 HOUSING WORKS LYMAN P 1412 PROSPECT AVENUE 28.0 162.0 \n", "399288 THOMAS MOTT OSBORNE M 549 EAST 171 STREET 75.0 141.0 \n", "\n", " stories fullval avland avtot \n", "0 5.0 87000 39150 774450 \n", "1 5.0 106000 47700 1153080 \n", "2 5.0 87000 39150 872640 \n", "3 2.0 314000 18840 95580 \n", "4 3.0 298000 17880 116520 \n", "... ... ... ... ... \n", "399284 3.0 422839 706 18814 \n", "399285 5.0 3346000 79200 1505700 \n", "399286 2.0 632000 13020 37920 \n", "399287 4.0 1174000 1362 76125 \n", "399288 1.0 188000 59850 84600 \n", "\n", "[399289 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.035.05.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.035.05.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.035.05.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.0100.02.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.0100.03.029800017880116520
................................................
399284234321941NaN20181AR3ALEMAN, CYNTHIA116 SURF DRIVE0.00.03.042283970618814
3992852333242NaN20182C73105 DECATUR ASSOCIAT3105 DECATUR AVENUE125.0119.05.03346000792001505700
3992862338538NaN20181B1HOGAN, RUTH345 EAST 236 STREET25.0100.02.06320001302037920
399287229636NaN20182BC1HOUSING WORKS LYMAN P1412 PROSPECT AVENUE28.0162.04.01174000136276125
3992882292833NaN20184Z9THOMAS MOTT OSBORNE M549 EAST 171 STREET75.0141.01.01880005985084600
\n", "

399289 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df3" } }, "metadata": {}, "execution_count": 69 } ], "source": [ "\n", "df3" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "coTjb3RM4Ez9", "outputId": "4b4b3e6e-9ac2-496f-f4c3-d04e47177c4b" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "0 3 2541 30 NaN 2021 NaN C1 \n", "1 3 2541 33 NaN 2021 NaN C1 \n", "2 3 2541 36 NaN 2021 NaN C1 \n", "3 3 2541 38 NaN 2021 NaN B9 \n", "4 3 2541 39 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "399284 2 3432 1941 NaN 2018 1A R3 \n", "399285 2 3332 42 NaN 2018 2 C7 \n", "399286 2 3385 38 NaN 2018 1 B1 \n", "399287 2 2963 6 NaN 2018 2B C1 \n", "399288 2 2928 33 NaN 2018 4 Z9 \n", "\n", " owner staddr ltfront ltdepth \\\n", "0 299 LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "1 293 LLC MCGUINNESS BOULEVARD 76.0 35.0 \n", "2 287A LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "3 JUSZCZAK, ANTONINA JAVA STREET 25.0 100.0 \n", "4 SAWICKI CZESLAW JAVA STREET 25.0 100.0 \n", "... ... ... ... ... \n", "399284 ALEMAN, CYNTHIA 116 SURF DRIVE 0.0 0.0 \n", "399285 3105 DECATUR ASSOCIAT 3105 DECATUR AVENUE 125.0 119.0 \n", "399286 HOGAN, RUTH 345 EAST 236 STREET 25.0 100.0 \n", "399287 HOUSING WORKS LYMAN P 1412 PROSPECT AVENUE 28.0 162.0 \n", "399288 THOMAS MOTT OSBORNE M 549 EAST 171 STREET 75.0 141.0 \n", "\n", " stories fullval avland avtot \n", "0 5.0 87000 39150 774450 \n", "1 5.0 106000 47700 1153080 \n", "2 5.0 87000 39150 872640 \n", "3 2.0 314000 18840 95580 \n", "4 3.0 298000 17880 116520 \n", "... ... ... ... ... \n", "399284 3.0 422839 706 18814 \n", "399285 5.0 3346000 79200 1505700 \n", "399286 2.0 632000 13020 37920 \n", "399287 4.0 1174000 1362 76125 \n", "399288 1.0 188000 59850 84600 \n", "\n", "[386557 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.035.05.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.035.05.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.035.05.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.0100.02.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.0100.03.029800017880116520
................................................
399284234321941NaN20181AR3ALEMAN, CYNTHIA116 SURF DRIVE0.00.03.042283970618814
3992852333242NaN20182C73105 DECATUR ASSOCIAT3105 DECATUR AVENUE125.0119.05.03346000792001505700
3992862338538NaN20181B1HOGAN, RUTH345 EAST 236 STREET25.0100.02.06320001302037920
399287229636NaN20182BC1HOUSING WORKS LYMAN P1412 PROSPECT AVENUE28.0162.04.01174000136276125
3992882292833NaN20184Z9THOMAS MOTT OSBORNE M549 EAST 171 STREET75.0141.01.01880005985084600
\n", "

386557 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df4" } }, "metadata": {}, "execution_count": 70 } ], "source": [ "# drop data without a full valuation\n", "df3['year'] = df3['year'].astype(int)\n", "df4 = df3.dropna(subset=['fullval'])\n", "\n", "\n", "# df4 = df3.dropna(subset=['staddr'])\n", "\n", "\n", "# Drop rows where FULLVAL is 0\n", "df4 = df4[df4['fullval'] != 0]" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "2Q8humcKhN9I" }, "outputs": [], "source": [ "if 'staddr' in df3.columns:\n", " # Drop rows with missing values in 'staddr'\n", " df4 = df3.dropna(subset=['staddr'])\n", " df4 = df3.dropna(subset=['fullval'])\n", "\n", " # Drop rows where FULLVAL is 0\n", " df4 = df4[df4['fullval'] != 0]\n", "else:\n", " print(\"'staddr' column not found in df3\")" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "M3ItjdrShIvO", "outputId": "8336ddd5-1e65-4e97-fd88-b4aaec0560a3" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "0 3 2541 30 NaN 2021 NaN C1 \n", "1 3 2541 33 NaN 2021 NaN C1 \n", "2 3 2541 36 NaN 2021 NaN C1 \n", "3 3 2541 38 NaN 2021 NaN B9 \n", "4 3 2541 39 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "399284 2 3432 1941 NaN 2018 1A R3 \n", "399285 2 3332 42 NaN 2018 2 C7 \n", "399286 2 3385 38 NaN 2018 1 B1 \n", "399287 2 2963 6 NaN 2018 2B C1 \n", "399288 2 2928 33 NaN 2018 4 Z9 \n", "\n", " owner staddr ltfront ltdepth \\\n", "0 299 LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "1 293 LLC MCGUINNESS BOULEVARD 76.0 35.0 \n", "2 287A LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "3 JUSZCZAK, ANTONINA JAVA STREET 25.0 100.0 \n", "4 SAWICKI CZESLAW JAVA STREET 25.0 100.0 \n", "... ... ... ... ... \n", "399284 ALEMAN, CYNTHIA 116 SURF DRIVE 0.0 0.0 \n", "399285 3105 DECATUR ASSOCIAT 3105 DECATUR AVENUE 125.0 119.0 \n", "399286 HOGAN, RUTH 345 EAST 236 STREET 25.0 100.0 \n", "399287 HOUSING WORKS LYMAN P 1412 PROSPECT AVENUE 28.0 162.0 \n", "399288 THOMAS MOTT OSBORNE M 549 EAST 171 STREET 75.0 141.0 \n", "\n", " stories fullval avland avtot \n", "0 5.0 87000 39150 774450 \n", "1 5.0 106000 47700 1153080 \n", "2 5.0 87000 39150 872640 \n", "3 2.0 314000 18840 95580 \n", "4 3.0 298000 17880 116520 \n", "... ... ... ... ... \n", "399284 3.0 422839 706 18814 \n", "399285 5.0 3346000 79200 1505700 \n", "399286 2.0 632000 13020 37920 \n", "399287 4.0 1174000 1362 76125 \n", "399288 1.0 188000 59850 84600 \n", "\n", "[386557 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.035.05.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.035.05.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.035.05.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.0100.02.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.0100.03.029800017880116520
................................................
399284234321941NaN20181AR3ALEMAN, CYNTHIA116 SURF DRIVE0.00.03.042283970618814
3992852333242NaN20182C73105 DECATUR ASSOCIAT3105 DECATUR AVENUE125.0119.05.03346000792001505700
3992862338538NaN20181B1HOGAN, RUTH345 EAST 236 STREET25.0100.02.06320001302037920
399287229636NaN20182BC1HOUSING WORKS LYMAN P1412 PROSPECT AVENUE28.0162.04.01174000136276125
3992882292833NaN20184Z9THOMAS MOTT OSBORNE M549 EAST 171 STREET75.0141.01.01880005985084600
\n", "

386557 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df4" } }, "metadata": {}, "execution_count": 72 } ], "source": [ "df4" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "46U2AegRyAx_" }, "outputs": [], "source": [ "from sklearn.compose import ColumnTransformer\n", "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.model_selection import train_test_split\n", "\n", "import pandas as pd\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import StandardScaler\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.svm import SVC\n", "from sklearn.metrics import accuracy_score, confusion_matrix\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 70 }, "id": "qoJ9TkUOvyGt", "outputId": "50326fd0-0a15-4bac-ba68-c03c5cf40b71" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "\"\\n# this to have address in a form comp can understand\\nfeature_transformer = ColumnTransformer(transformers=[\\n ('onehot', OneHotEncoder(handle_unknown='ignore'), ['staddr'])\\n], remainder='passthrough')\\n\\n\\npipeline = Pipeline([\\n ('transformer', feature_transformer),\\n ('scaler', StandardScaler(with_mean=False)) # Set with_mean to False\\n])\\n\\n# Split dataset\\nX = df4[['year', 'staddr']]\\ny = df4['fullval']\\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)\\n\\n# Apply transformations using the pipeline\\nX_train = pipeline.fit_transform(X_train)\\nX_test = pipeline.transform(X_test)\\n\"" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "string" } }, "metadata": {}, "execution_count": 74 } ], "source": [ "'''\n", "# this to have address in a form comp can understand\n", "feature_transformer = ColumnTransformer(transformers=[\n", " ('onehot', OneHotEncoder(handle_unknown='ignore'), ['staddr'])\n", "], remainder='passthrough')\n", "\n", "\n", "pipeline = Pipeline([\n", " ('transformer', feature_transformer),\n", " ('scaler', StandardScaler(with_mean=False)) # Set with_mean to False\n", "])\n", "\n", "# Split dataset\n", "X = df4[['year', 'staddr']]\n", "y = df4['fullval']\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)\n", "\n", "# Apply transformations using the pipeline\n", "X_train = pipeline.fit_transform(X_train)\n", "X_test = pipeline.transform(X_test)\n", "'''" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "JAZaJz4e4HmY" }, "outputs": [], "source": [ "from sklearn.ensemble import RandomForestRegressor\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.svm import SVR\n", "from sklearn.metrics import mean_squared_error, r2_score" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "dmrATUHjyx5k" }, "outputs": [], "source": [ "# Define the regressors to use\n", "regressors = {\n", " 'Random Forest Regressor': RandomForestRegressor(n_estimators=100),\n", " 'Linear Regression': LinearRegression(),\n", " 'Support Vector Regressor': SVR()\n", "}" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "L6761eQ1gsNz", "outputId": "a35bd72e-d3dc-4052-b1d0-8e86a8af1946" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "0 3 2541 30 NaN 2021 NaN C1 \n", "1 3 2541 33 NaN 2021 NaN C1 \n", "2 3 2541 36 NaN 2021 NaN C1 \n", "3 3 2541 38 NaN 2021 NaN B9 \n", "4 3 2541 39 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "399284 2 3432 1941 NaN 2018 1A R3 \n", "399285 2 3332 42 NaN 2018 2 C7 \n", "399286 2 3385 38 NaN 2018 1 B1 \n", "399287 2 2963 6 NaN 2018 2B C1 \n", "399288 2 2928 33 NaN 2018 4 Z9 \n", "\n", " owner staddr ltfront ltdepth \\\n", "0 299 LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "1 293 LLC MCGUINNESS BOULEVARD 76.0 35.0 \n", "2 287A LLC MCGUINNESS BOULEVARD 62.0 35.0 \n", "3 JUSZCZAK, ANTONINA JAVA STREET 25.0 100.0 \n", "4 SAWICKI CZESLAW JAVA STREET 25.0 100.0 \n", "... ... ... ... ... \n", "399284 ALEMAN, CYNTHIA 116 SURF DRIVE 0.0 0.0 \n", "399285 3105 DECATUR ASSOCIAT 3105 DECATUR AVENUE 125.0 119.0 \n", "399286 HOGAN, RUTH 345 EAST 236 STREET 25.0 100.0 \n", "399287 HOUSING WORKS LYMAN P 1412 PROSPECT AVENUE 28.0 162.0 \n", "399288 THOMAS MOTT OSBORNE M 549 EAST 171 STREET 75.0 141.0 \n", "\n", " stories fullval avland avtot \n", "0 5.0 87000 39150 774450 \n", "1 5.0 106000 47700 1153080 \n", "2 5.0 87000 39150 872640 \n", "3 2.0 314000 18840 95580 \n", "4 3.0 298000 17880 116520 \n", "... ... ... ... ... \n", "399284 3.0 422839 706 18814 \n", "399285 5.0 3346000 79200 1505700 \n", "399286 2.0 632000 13020 37920 \n", "399287 4.0 1174000 1362 76125 \n", "399288 1.0 188000 59850 84600 \n", "\n", "[386557 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.035.05.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.035.05.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.035.05.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.0100.02.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.0100.03.029800017880116520
................................................
399284234321941NaN20181AR3ALEMAN, CYNTHIA116 SURF DRIVE0.00.03.042283970618814
3992852333242NaN20182C73105 DECATUR ASSOCIAT3105 DECATUR AVENUE125.0119.05.03346000792001505700
3992862338538NaN20181B1HOGAN, RUTH345 EAST 236 STREET25.0100.02.06320001302037920
399287229636NaN20182BC1HOUSING WORKS LYMAN P1412 PROSPECT AVENUE28.0162.04.01174000136276125
3992882292833NaN20184Z9THOMAS MOTT OSBORNE M549 EAST 171 STREET75.0141.01.01880005985084600
\n", "

386557 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df4" } }, "metadata": {}, "execution_count": 77 } ], "source": [ "df4" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "ceJMjQjp436e" }, "outputs": [], "source": [ "# Because each borough is likely to exhibit different growth charatesitscs I have deceided to split the dataset by borough\n", "\n", "# Split the dataframe based on the 'boro' column\n", "Manhattandf = df4[df4['boro'] == 1]\n", "Bronxdf = df4[df4['boro'] == 2]\n", "Brooklyndf = df4[df4['boro'] == 3]\n", "Queensdf = df4[df4['boro'] == 4]\n", "StatenIslanddf = df4[df4['boro'] == 5]\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "4nIwI_TJyHh5", "outputId": "6663bb51-c167-4a4c-d6af-74809a7d7982" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "6524 1 1841 13 NaN 2021 NaN B3 \n", "8244 1 1841 14 NaN 2021 NaN A5 \n", "8323 1 1841 15 NaN 2021 NaN A4 \n", "8477 1 1841 16 NaN 2021 NaN A9 \n", "9218 1 1841 17 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "398253 1 2225 30 NaN 2018 2 C1 \n", "398254 1 2215 356 NaN 2018 1 B1 \n", "398264 1 2202 9 NaN 2018 2 D4 \n", "398265 1 2164 44 NaN 2018 2 C7 \n", "398283 1 2175 63 NaN 2018 4 K1 \n", "\n", " owner staddr ltfront ltdepth \\\n", "6524 GARLAND E WOOD WEST 105 STREET 17.58 75.00 \n", "8244 NOYES, ELIZABETH MANHATTAN AVENUE 16.67 75.00 \n", "8323 GOINGS, MATTIE M MANHATTAN AVENUE 16.33 86.83 \n", "8477 GLENN GULLICKSON, TIMOTHY MANHATTAN AVENUE 16.33 86.83 \n", "9218 RAYMOND RECHT MANHATTAN AVENUE 17.00 70.00 \n", "... ... ... ... ... \n", "398253 M & N MANAGEMENT CORP 168 SHERMAN AVENUE 50.00 150.00 \n", "398254 PONDICHI CONSTANTIN 10 VAN CORLEAR PLACE 28.00 133.00 \n", "398264 420 WEST 206 STREET O 416 WEST 206 STREET 200.00 99.00 \n", "398265 175 REALTY ASSOCIATES 4316 BROADWAY 37.00 103.00 \n", "398283 PARTNERS 2004 LLC 180 DYCKMAN STREET 50.00 100.00 \n", "\n", " stories fullval avland avtot \n", "6524 3.0 1107000 66420 175440 \n", "8244 3.0 226000 13560 219780 \n", "8323 3.0 1179000 70740 227580 \n", "8477 3.0 1782000 106920 256260 \n", "9218 3.0 1611000 96660 219300 \n", "... ... ... ... ... \n", "398253 5.0 1584000 59850 712800 \n", "398254 2.0 655000 23674 34006 \n", "398264 6.0 6011000 193950 2704950 \n", "398265 5.0 2122000 59400 954900 \n", "398283 1.0 2320000 201600 1044000 \n", "\n", "[155480 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
65241184113NaN2021NaNB3GARLAND E WOODWEST 105 STREET17.5875.003.0110700066420175440
82441184114NaN2021NaNA5NOYES, ELIZABETHMANHATTAN AVENUE16.6775.003.022600013560219780
83231184115NaN2021NaNA4GOINGS, MATTIE MMANHATTAN AVENUE16.3386.833.0117900070740227580
84771184116NaN2021NaNA9GLENN GULLICKSON, TIMOTHYMANHATTAN AVENUE16.3386.833.01782000106920256260
92181184117NaN2021NaNB9RAYMOND RECHTMANHATTAN AVENUE17.0070.003.0161100096660219300
................................................
3982531222530NaN20182C1M & N MANAGEMENT CORP168 SHERMAN AVENUE50.00150.005.0158400059850712800
39825412215356NaN20181B1PONDICHI CONSTANTIN10 VAN CORLEAR PLACE28.00133.002.06550002367434006
398264122029NaN20182D4420 WEST 206 STREET O416 WEST 206 STREET200.0099.006.060110001939502704950
3982651216444NaN20182C7175 REALTY ASSOCIATES4316 BROADWAY37.00103.005.0212200059400954900
3982831217563NaN20184K1PARTNERS 2004 LLC180 DYCKMAN STREET50.00100.001.023200002016001044000
\n", "

155480 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "Manhattandf" } }, "metadata": {}, "execution_count": 79 } ], "source": [ "Manhattandf" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 567 }, "id": "DcP_VRE47St7", "outputId": "ac3c0973-9379-49f4-8316-d055df2a9974" }, "outputs": [ { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": {} }, { "output_type": "stream", "name": "stdout", "text": [ "\n", " {'Manhattan': (2010, 2021), 'Bronx': (2010, 2021), 'Brooklyn': (2010, 2022), 'Queens': (2010, 2018), 'Staten Island': (2010, 2018)}\n" ] } ], "source": [ "# create a dictionary with the borough names and the number of rows in each dataframe\n", "borough_counts = {\n", " 'Manhattan': len(Manhattandf),\n", " 'Bronx': len(Bronxdf),\n", " 'Brooklyn': len(Brooklyndf),\n", " 'Queens': len(Queensdf),\n", " 'Staten Island': len(StatenIslanddf)\n", "}\n", "\n", "# plot this information using a bar graph\n", "plt.bar(borough_counts.keys(), borough_counts.values())\n", "plt.xlabel('Borough')\n", "plt.ylabel('Number of Data Points')\n", "plt.title('Data Points per Borough')\n", "plt.xticks(rotation=45)\n", "plt.show()\n", "\n", "\n", "# date ranges in each dataframe\n", "date_ranges = {\n", " 'Manhattan': (Manhattandf['year'].min(), Manhattandf['year'].max()),\n", " 'Bronx': (Bronxdf['year'].min(), Bronxdf['year'].max()),\n", " 'Brooklyn': (Brooklyndf['year'].min(), Brooklyndf['year'].max()),\n", " 'Queens': (Queensdf['year'].min(), Queensdf['year'].max()),\n", " 'Staten Island': (StatenIslanddf['year'].min(), StatenIslanddf['year'].max())\n", "}\n", "\n", "print(\"\\n\", date_ranges)\n", "\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "BMu6STSB8hzk", "outputId": "6846abee-fffc-4b81-8bdf-89e3b65a4833" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "6524 1 1841 13 NaN 2021 NaN B3 \n", "8244 1 1841 14 NaN 2021 NaN A5 \n", "8323 1 1841 15 NaN 2021 NaN A4 \n", "8477 1 1841 16 NaN 2021 NaN A9 \n", "9218 1 1841 17 NaN 2021 NaN B9 \n", "... ... ... ... ... ... ... ... \n", "398253 1 2225 30 NaN 2018 2 C1 \n", "398254 1 2215 356 NaN 2018 1 B1 \n", "398264 1 2202 9 NaN 2018 2 D4 \n", "398265 1 2164 44 NaN 2018 2 C7 \n", "398283 1 2175 63 NaN 2018 4 K1 \n", "\n", " owner staddr ltfront ltdepth \\\n", "6524 GARLAND E WOOD WEST 105 STREET 17.58 75.00 \n", "8244 NOYES, ELIZABETH MANHATTAN AVENUE 16.67 75.00 \n", "8323 GOINGS, MATTIE M MANHATTAN AVENUE 16.33 86.83 \n", "8477 GLENN GULLICKSON, TIMOTHY MANHATTAN AVENUE 16.33 86.83 \n", "9218 RAYMOND RECHT MANHATTAN AVENUE 17.00 70.00 \n", "... ... ... ... ... \n", "398253 M & N MANAGEMENT CORP 168 SHERMAN AVENUE 50.00 150.00 \n", "398254 PONDICHI CONSTANTIN 10 VAN CORLEAR PLACE 28.00 133.00 \n", "398264 420 WEST 206 STREET O 416 WEST 206 STREET 200.00 99.00 \n", "398265 175 REALTY ASSOCIATES 4316 BROADWAY 37.00 103.00 \n", "398283 PARTNERS 2004 LLC 180 DYCKMAN STREET 50.00 100.00 \n", "\n", " stories fullval avland avtot \n", "6524 3.0 1107000 66420 175440 \n", "8244 3.0 226000 13560 219780 \n", "8323 3.0 1179000 70740 227580 \n", "8477 3.0 1782000 106920 256260 \n", "9218 3.0 1611000 96660 219300 \n", "... ... ... ... ... \n", "398253 5.0 1584000 59850 712800 \n", "398254 2.0 655000 23674 34006 \n", "398264 6.0 6011000 193950 2704950 \n", "398265 5.0 2122000 59400 954900 \n", "398283 1.0 2320000 201600 1044000 \n", "\n", "[155480 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
65241184113NaN2021NaNB3GARLAND E WOODWEST 105 STREET17.5875.003.0110700066420175440
82441184114NaN2021NaNA5NOYES, ELIZABETHMANHATTAN AVENUE16.6775.003.022600013560219780
83231184115NaN2021NaNA4GOINGS, MATTIE MMANHATTAN AVENUE16.3386.833.0117900070740227580
84771184116NaN2021NaNA9GLENN GULLICKSON, TIMOTHYMANHATTAN AVENUE16.3386.833.01782000106920256260
92181184117NaN2021NaNB9RAYMOND RECHTMANHATTAN AVENUE17.0070.003.0161100096660219300
................................................
3982531222530NaN20182C1M & N MANAGEMENT CORP168 SHERMAN AVENUE50.00150.005.0158400059850712800
39825412215356NaN20181B1PONDICHI CONSTANTIN10 VAN CORLEAR PLACE28.00133.002.06550002367434006
398264122029NaN20182D4420 WEST 206 STREET O416 WEST 206 STREET200.0099.006.060110001939502704950
3982651216444NaN20182C7175 REALTY ASSOCIATES4316 BROADWAY37.00103.005.0212200059400954900
3982831217563NaN20184K1PARTNERS 2004 LLC180 DYCKMAN STREET50.00100.001.023200002016001044000
\n", "

155480 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "Manhattandf" } }, "metadata": {}, "execution_count": 81 } ], "source": [ "Manhattandf" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "KtFSAiTH8pSF", "outputId": "19ca7a54-c015-415a-d17d-e2e36a794e27" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl owner \\\n", "0 3 2541 30 NaN 2021 NaN C1 299 LLC \n", "1 3 2541 33 NaN 2021 NaN C1 293 LLC \n", "2 3 2541 36 NaN 2021 NaN C1 287A LLC \n", "3 3 2541 38 NaN 2021 NaN B9 JUSZCZAK, ANTONINA \n", "4 3 2541 39 NaN 2021 NaN B9 SAWICKI CZESLAW \n", "... ... ... ... ... ... ... ... ... \n", "229673 3 8932 682 NaN 2011 1B V0 WILLIAM GLEASON \n", "229674 3 8946 990 NaN 2011 1B V0 FRANCES NELSON \n", "229675 3 8955 204 NaN 2011 1B V0 E SARUBBI \n", "311669 3 865 1 NaN 2018 4 M4 THE MISSNRY SOC HOLY \n", "313472 3 865 1 NaN 2018 4 M4 THE MISSNRY SOC HOLY \n", "\n", " staddr ltfront ltdepth stories fullval avland \\\n", "0 MCGUINNESS BOULEVARD 62.0 35.0 5.0 87000 39150 \n", "1 MCGUINNESS BOULEVARD 76.0 35.0 5.0 106000 47700 \n", "2 MCGUINNESS BOULEVARD 62.0 35.0 5.0 87000 39150 \n", "3 JAVA STREET 25.0 100.0 2.0 314000 18840 \n", "4 JAVA STREET 25.0 100.0 3.0 298000 17880 \n", "... ... ... ... ... ... ... \n", "229673 DICTUM COURT 34.0 52.0 NaN 143000 2073 \n", "229674 DICTUM COURT 34.0 52.0 NaN 143000 2593 \n", "229675 CELESTE COURT 34.0 52.0 NaN 143000 1860 \n", "311669 5901 6 AVENUE 200.0 100.0 4.0 3069000 253350 \n", "313472 5901 6 AVENUE 200.0 100.0 4.0 3069000 253350 \n", "\n", " avtot \n", "0 774450 \n", "1 1153080 \n", "2 872640 \n", "3 95580 \n", "4 116520 \n", "... ... \n", "229673 2073 \n", "229674 2593 \n", "229675 1860 \n", "311669 1381050 \n", "313472 1381050 \n", "\n", "[197517 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
03254130NaN2021NaNC1299 LLCMCGUINNESS BOULEVARD62.035.05.08700039150774450
13254133NaN2021NaNC1293 LLCMCGUINNESS BOULEVARD76.035.05.0106000477001153080
23254136NaN2021NaNC1287A LLCMCGUINNESS BOULEVARD62.035.05.08700039150872640
33254138NaN2021NaNB9JUSZCZAK, ANTONINAJAVA STREET25.0100.02.03140001884095580
43254139NaN2021NaNB9SAWICKI CZESLAWJAVA STREET25.0100.03.029800017880116520
................................................
22967338932682NaN20111BV0WILLIAM GLEASONDICTUM COURT34.052.0NaN14300020732073
22967438946990NaN20111BV0FRANCES NELSONDICTUM COURT34.052.0NaN14300025932593
22967538955204NaN20111BV0E SARUBBICELESTE COURT34.052.0NaN14300018601860
31166938651NaN20184M4THE MISSNRY SOC HOLY5901 6 AVENUE200.0100.04.030690002533501381050
31347238651NaN20184M4THE MISSNRY SOC HOLY5901 6 AVENUE200.0100.04.030690002533501381050
\n", "

197517 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "Brooklyndf" } }, "metadata": {}, "execution_count": 82 } ], "source": [ "Brooklyndf" ] }, { "cell_type": "code", "source": [ "# temp\n", "\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.metrics import mean_squared_error, r2_score\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.ensemble import RandomForestRegressor\n", "from xgboost import XGBRegressor\n", "\n", "\n", "# Feature Selection and Engineering\n", "df_combined = Manhattandf\n", "\n", "# Removed staddr as we already have boro and block numbers so being further specific will serve little purpose\n", "# features_to_use = ['boro', 'block', 'lot', 'taxclass', 'bldgcl', 'staddr', 'ltfront', 'ltdepth', 'stories']\n", "features_to_use = ['block', 'taxclass', 'bldgcl', 'ltfront', 'ltdepth', 'stories', 'year']\n", "target = 'fullval'\n", "\n", "# Data Cleaning\n", "df_combined.dropna(subset=[target], inplace=True)\n", "\n", "# Convert categorical features to string\n", "# categorical_features = ['boro', 'taxclass', 'bldgcl', 'staddr'] # Assuming 'staddr' is categorical\n", "categorical_features = ['taxclass', 'bldgcl']\n", "for col in categorical_features:\n", " df_combined[col] = df_combined[col].astype(str)\n", "\n", "# Categorize features\n", "numerical_features = [f for f in features_to_use if f not in categorical_features]\n", "\n", "# Preprocessor for handling both categorical and numerical features\n", "preprocessor = ColumnTransformer(\n", " transformers=[\n", " ('num', Pipeline([\n", " ('imputer', SimpleImputer(strategy='mean')), # or 'median'\n", " ('scaler', StandardScaler())\n", " ]), numerical_features),\n", " ('cat', Pipeline([\n", " ('imputer', SimpleImputer(strategy='constant', fill_value='missing')),\n", " ('encoder', OneHotEncoder(handle_unknown='ignore'))\n", " ]), categorical_features)\n", " ]\n", ")\n", "'''\n", "# Linear Regression Model\n", "# Pipeline with model\n", "# Yielded .44 with 120,000 Values\n", "# Yielded .22 with 240,000 Values\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', LinearRegression())\n", "])\n", "'''\n", "'''\n", "# This model takes too long to load\n", "# Random Forest Regressor Model\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', RandomForestRegressor(n_estimators=100, random_state=42))\n", "])\n", "'''\n", "\n", "\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', XGBRegressor(objective='reg:squarederror', n_estimators=100, learning_rate=0.3, random_state=42))\n", "])\n", "\n", "\n", "# Model preparation\n", "X = df_combined[features_to_use]\n", "y = df_combined[target]\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n", "\n", "# Train the model\n", "pipeline.fit(X_train, y_train)\n", "\n", "# Predictions and Evaluation\n", "y_pred = pipeline.predict(X_test)\n", "mse = mean_squared_error(y_test, y_pred)\n", "r2 = r2_score(y_test, y_pred)\n", "\n", "print(f'Mean Squared Error: {mse}')\n", "print(f'R-squared: {r2}')\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Z5_uEDHdFEEU", "outputId": "efd258d5-d869-493f-c3b9-1f99c6926a3b" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ ":23: SettingWithCopyWarning: \n", "A value is trying to be set on a copy of a slice from a DataFrame\n", "\n", "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", " df_combined.dropna(subset=[target], inplace=True)\n", ":29: SettingWithCopyWarning: \n", "A value is trying to be set on a copy of a slice from a DataFrame.\n", "Try using .loc[row_indexer,col_indexer] = value instead\n", "\n", "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", " df_combined[col] = df_combined[col].astype(str)\n" ] }, { "output_type": "stream", "name": "stdout", "text": [ "Mean Squared Error: 108184632975999.38\n", "R-squared: 0.7479928419732339\n" ] } ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 300 }, "id": "UUFlHrzE95r8", "outputId": "2016278a-6f7f-48f2-b668-f4374af94b4b" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " Metric Manhattan Brooklyn\n", "0 Mean Full Value 3.624771e+06 267494.733552\n", "1 Median Full Value 4.224895e+05 221000.000000\n", "2 Mean AV Land 3.066454e+05 40435.405768\n", "3 Median AV Land 2.521200e+04 15022.000000\n", "4 Mean AV Total 1.449849e+06 180985.322139\n", "5 Median AV Total 1.654295e+05 62100.000000\n", "6 Mean Stories 1.837571e+01 2.732021\n", "7 Median Stories 1.300000e+01 2.000000" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MetricManhattanBrooklyn
0Mean Full Value3.624771e+06267494.733552
1Median Full Value4.224895e+05221000.000000
2Mean AV Land3.066454e+0540435.405768
3Median AV Land2.521200e+0415022.000000
4Mean AV Total1.449849e+06180985.322139
5Median AV Total1.654295e+0562100.000000
6Mean Stories1.837571e+012.732021
7Median Stories1.300000e+012.000000
\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "metrics_df", "summary": "{\n \"name\": \"metrics_df\",\n \"rows\": 8,\n \"fields\": [\n {\n \"column\": \"Metric\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 8,\n \"samples\": [\n \"Median Full Value\",\n \"Median AV Total\",\n \"Mean Full Value\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Manhattan\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1256407.2290299018,\n \"min\": 13.0,\n \"max\": 3624771.173540005,\n \"num_unique_values\": 8,\n \"samples\": [\n 422489.5,\n 165429.5,\n 3624771.173540005\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Brooklyn\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 107860.87934620428,\n \"min\": 2.0,\n \"max\": 267494.7335520487,\n \"num_unique_values\": 8,\n \"samples\": [\n 221000.0,\n 62100.0,\n 267494.7335520487\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}" } }, "metadata": {}, "execution_count": 84 } ], "source": [ "# Trying to figure out what qualities is leading to Brooklyn having a far lower accuracy score than manhattan\n", "# So far my current explanation is that Brooklyn buildings are mostly residential and the accuracy to predict the prices in very differing areas makes it difficult for the model to predict accurately the value of\n", "\n", "metrics = {\n", " 'Metric': ['Mean Full Value', 'Median Full Value', 'Mean AV Land', 'Median AV Land', 'Mean AV Total', 'Median AV Total', 'Mean Stories', 'Median Stories'],\n", " 'Manhattan': [\n", " Manhattandf['fullval'].mean(),\n", " Manhattandf['fullval'].median(),\n", " Manhattandf['avland'].mean(),\n", " Manhattandf['avland'].median(),\n", " Manhattandf['avtot'].mean(),\n", " Manhattandf['avtot'].median(),\n", " Manhattandf['stories'].mean(),\n", " Manhattandf['stories'].median(),\n", " ],\n", " 'Brooklyn': [\n", " Brooklyndf['fullval'].mean(),\n", " Brooklyndf['fullval'].median(),\n", " Brooklyndf['avland'].mean(),\n", " Brooklyndf['avland'].median(),\n", " Brooklyndf['avtot'].mean(),\n", " Brooklyndf['avtot'].median(),\n", " Brooklyndf['stories'].mean(),\n", " Brooklyndf['stories'].median(),\n", " ]\n", "}\n", "\n", "# Converting the metrics to a DataFrame for display\n", "metrics_df = pd.DataFrame(metrics)\n", "metrics_df" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "dx2r5Qev-49Q", "outputId": "7677a597-002e-4392-8351-fc33991ca5b1" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(197517, 91637)" ] }, "metadata": {}, "execution_count": 85 } ], "source": [ "# Attempting to remove any extreme outliers from the dataset\n", "\n", "def remove_outliers(df, column):\n", " mean_val = df[column].mean()\n", " # Setting the deviation limit to 60% of the mean\n", " deviation_limit = 0.6 * mean_val\n", " lower_bound = mean_val - deviation_limit\n", " upper_bound = mean_val + deviation_limit\n", " # Filtering the DataFrame to only include values within the specified range\n", " return df[(df[column] >= lower_bound) & (df[column] <= upper_bound)]\n", "\n", "# Applying the outlier removal function to the relevant columns in the Manhattan and Brooklyn dataframes\n", "columns_to_check = ['fullval', 'avland', 'avtot', 'stories']\n", "\n", "Brooklyndf_cleaned = Brooklyndf.copy()\n", "\n", "for column in columns_to_check:\n", " Brooklyndf_cleaned = remove_outliers(Brooklyndf_cleaned, column)\n", "\n", "# Display the size of the dataframes before and after outlier removal\n", "original_sizes = len(Brooklyndf)\n", "cleaned_sizes = len(Brooklyndf_cleaned)\n", "\n", "original_sizes, cleaned_sizes" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 300 }, "id": "WgSs90Ixvs3s", "outputId": "42bc346b-9bb1-4f67-b005-4104cbd03cfd" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " Metric Brooklyn\n", "0 Mean Full Value 265926.204044\n", "1 Median Full Value 253000.000000\n", "2 Mean AV Land 15954.862697\n", "3 Median AV Land 15180.000000\n", "4 Mean AV Total 56856.716588\n", "5 Median AV Total 56460.000000\n", "6 Mean Stories 2.116288\n", "7 Median Stories 2.000000" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MetricBrooklyn
0Mean Full Value265926.204044
1Median Full Value253000.000000
2Mean AV Land15954.862697
3Median AV Land15180.000000
4Mean AV Total56856.716588
5Median AV Total56460.000000
6Mean Stories2.116288
7Median Stories2.000000
\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "metrics_df", "summary": "{\n \"name\": \"metrics_df\",\n \"rows\": 8,\n \"fields\": [\n {\n \"column\": \"Metric\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 8,\n \"samples\": [\n \"Median Full Value\",\n \"Median AV Total\",\n \"Mean Full Value\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Brooklyn\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 111240.86529315186,\n \"min\": 2.0,\n \"max\": 265926.20404421794,\n \"num_unique_values\": 8,\n \"samples\": [\n 253000.0,\n 56460.0,\n 265926.20404421794\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}" } }, "metadata": {}, "execution_count": 86 } ], "source": [ "metrics = {\n", " 'Metric': ['Mean Full Value', 'Median Full Value', 'Mean AV Land', 'Median AV Land', 'Mean AV Total', 'Median AV Total', 'Mean Stories', 'Median Stories'],\n", " 'Brooklyn': [\n", " Brooklyndf_cleaned['fullval'].mean(),\n", " Brooklyndf_cleaned['fullval'].median(),\n", " Brooklyndf_cleaned['avland'].mean(),\n", " Brooklyndf_cleaned['avland'].median(),\n", " Brooklyndf_cleaned['avtot'].mean(),\n", " Brooklyndf_cleaned['avtot'].median(),\n", " Brooklyndf_cleaned['stories'].mean(),\n", " Brooklyndf_cleaned['stories'].median(),\n", " ]\n", "}\n", "\n", "# Converting the metrics to a DataFrame for display\n", "metrics_df = pd.DataFrame(metrics)\n", "metrics_df" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 616 }, "id": "2lHJn-B505IE", "outputId": "db14c7e0-a59d-4fb6-f637-fdba05cb653a" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "481576.25\n", "369509\n", "603541\n" ] }, { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": {} } ], "source": [ "# Used this to create a bell curve and discovered that there are outliers in price that are messing with certain predictions resulting in negative values\n", "# Negative values originate from the fact that there is data supporting a very low unrealistic price for a given year in a block that messes the prediction for other buildings in the same block\n", "# This was used for testing, this has no effect on the Manhattandf and the df referred here isnt used later\n", "# Information gathered here was used for other code\n", "\n", "import pandas as pd\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "from scipy.stats import norm\n", "\n", "df = Manhattandf\n", "\n", "block_input = 16\n", "year_input = 2017\n", "\n", "# Filter the DataFrame for the specified block and year\n", "filtered_df = df[(df['block'] == block_input) & (df['year'] == year_input)]\n", "\n", "\n", "prices = filtered_df['fullval'].dropna() # Drop NaN values\n", "\n", "mean = prices.mean()\n", "std = prices.std()\n", "\n", "# Generate a range of price values for the bell curve\n", "price_range = np.linspace(prices.min(), prices.max(), 100)\n", "\n", "bell_curve = norm.pdf(price_range, mean, std)\n", "\n", "print(mean)\n", "print(prices.min())\n", "print(prices.max())\n", "\n", "plt.figure(figsize=(10, 6))\n", "plt.plot(price_range, bell_curve, color='blue')\n", "plt.title(f'Price Distribution of Buildings in Block {block_input} for the Year {year_input}')\n", "plt.xlabel('Price')\n", "plt.ylabel('Density')\n", "plt.grid(True)\n", "plt.show()\n" ] }, { "cell_type": "code", "source": [ "#temp\n", "\n", "\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.metrics import mean_squared_error, r2_score\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.ensemble import RandomForestRegressor\n", "from xgboost import XGBRegressor\n", "\n", "\n", "# Feature Selection and Engineering\n", "df_combined = Manhattandf\n", "\n", "# Removed staddr as we already have boro and block numbers so being further specific will serve little purpose\n", "# features_to_use = ['boro', 'block', 'lot', 'taxclass', 'bldgcl', 'staddr', 'ltfront', 'ltdepth', 'stories']\n", "features_to_use = ['block', 'taxclass', 'bldgcl', 'ltfront', 'ltdepth', 'stories', 'year']\n", "target = 'fullval'\n", "\n", "# Data Cleaning\n", "df_combined.dropna(subset=[target], inplace=True)\n", "\n", "# Convert categorical features to string\n", "# categorical_features = ['boro', 'taxclass', 'bldgcl', 'staddr'] # Assuming 'staddr' is categorical\n", "categorical_features = ['taxclass', 'bldgcl']\n", "for col in categorical_features:\n", " df_combined[col] = df_combined[col].astype(str)\n", "\n", "# Categorize features\n", "numerical_features = [f for f in features_to_use if f not in categorical_features]\n", "\n", "# Preprocessor for handling both categorical and numerical features\n", "preprocessor = ColumnTransformer(\n", " transformers=[\n", " ('num', Pipeline([\n", " ('imputer', SimpleImputer(strategy='mean')), # or 'median'\n", " ('scaler', StandardScaler())\n", " ]), numerical_features),\n", " ('cat', Pipeline([\n", " ('imputer', SimpleImputer(strategy='constant', fill_value='missing')),\n", " ('encoder', OneHotEncoder(handle_unknown='ignore'))\n", " ]), categorical_features)\n", " ]\n", ")\n", "'''\n", "# Linear Regression Model\n", "# Pipeline with model\n", "# Yielded .44 with 120,000 Values\n", "# Yielded .22 with 240,000 Values\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', LinearRegression())\n", "])\n", "'''\n", "'''\n", "# This model takes too long to load\n", "# Random Forest Regressor Model\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', RandomForestRegressor(n_estimators=100, random_state=42))\n", "])\n", "'''\n", "\n", "\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', XGBRegressor(objective='reg:squarederror', n_estimators=100, learning_rate=0.3, random_state=42))\n", "])\n", "\n", "\n", "# Model preparation\n", "X = df_combined[features_to_use]\n", "y = df_combined[target]\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n", "\n", "# Train the model\n", "pipeline.fit(X_train, y_train)\n", "\n", "# Predictions and Evaluation\n", "y_pred = pipeline.predict(X_test)\n", "mse = mean_squared_error(y_test, y_pred)\n", "r2 = r2_score(y_test, y_pred)\n", "\n", "print(f'Mean Squared Error: {mse}')\n", "print(f'R-squared: {r2}')\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "lrmPONdYE2_5", "outputId": "c6e4a024-bdef-4e9c-abf7-d422c2a5597c" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ ":24: SettingWithCopyWarning: \n", "A value is trying to be set on a copy of a slice from a DataFrame\n", "\n", "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", " df_combined.dropna(subset=[target], inplace=True)\n", ":30: SettingWithCopyWarning: \n", "A value is trying to be set on a copy of a slice from a DataFrame.\n", "Try using .loc[row_indexer,col_indexer] = value instead\n", "\n", "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", " df_combined[col] = df_combined[col].astype(str)\n" ] }, { "output_type": "stream", "name": "stdout", "text": [ "Mean Squared Error: 108184632975999.38\n", "R-squared: 0.7479928419732339\n" ] } ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 424 }, "id": "m3shi1MYg2id", "outputId": "270c22cd-d8d0-41cd-8696-1a3e63aa5ed1" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "6524 1 1841 13 NaN 2021 nan B3 \n", "8244 1 1841 14 NaN 2021 nan A5 \n", "8323 1 1841 15 NaN 2021 nan A4 \n", "8477 1 1841 16 NaN 2021 nan A9 \n", "9218 1 1841 17 NaN 2021 nan B9 \n", "... ... ... ... ... ... ... ... \n", "398253 1 2225 30 NaN 2018 2 C1 \n", "398254 1 2215 356 NaN 2018 1 B1 \n", "398264 1 2202 9 NaN 2018 2 D4 \n", "398265 1 2164 44 NaN 2018 2 C7 \n", "398283 1 2175 63 NaN 2018 4 K1 \n", "\n", " owner staddr ltfront ltdepth \\\n", "6524 GARLAND E WOOD WEST 105 STREET 17.58 75.00 \n", "8244 NOYES, ELIZABETH MANHATTAN AVENUE 16.67 75.00 \n", "8323 GOINGS, MATTIE M MANHATTAN AVENUE 16.33 86.83 \n", "8477 GLENN GULLICKSON, TIMOTHY MANHATTAN AVENUE 16.33 86.83 \n", "9218 RAYMOND RECHT MANHATTAN AVENUE 17.00 70.00 \n", "... ... ... ... ... \n", "398253 M & N MANAGEMENT CORP 168 SHERMAN AVENUE 50.00 150.00 \n", "398254 PONDICHI CONSTANTIN 10 VAN CORLEAR PLACE 28.00 133.00 \n", "398264 420 WEST 206 STREET O 416 WEST 206 STREET 200.00 99.00 \n", "398265 175 REALTY ASSOCIATES 4316 BROADWAY 37.00 103.00 \n", "398283 PARTNERS 2004 LLC 180 DYCKMAN STREET 50.00 100.00 \n", "\n", " stories fullval avland avtot \n", "6524 3.0 1107000 66420 175440 \n", "8244 3.0 226000 13560 219780 \n", "8323 3.0 1179000 70740 227580 \n", "8477 3.0 1782000 106920 256260 \n", "9218 3.0 1611000 96660 219300 \n", "... ... ... ... ... \n", "398253 5.0 1584000 59850 712800 \n", "398254 2.0 655000 23674 34006 \n", "398264 6.0 6011000 193950 2704950 \n", "398265 5.0 2122000 59400 954900 \n", "398283 1.0 2320000 201600 1044000 \n", "\n", "[155480 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
65241184113NaN2021nanB3GARLAND E WOODWEST 105 STREET17.5875.003.0110700066420175440
82441184114NaN2021nanA5NOYES, ELIZABETHMANHATTAN AVENUE16.6775.003.022600013560219780
83231184115NaN2021nanA4GOINGS, MATTIE MMANHATTAN AVENUE16.3386.833.0117900070740227580
84771184116NaN2021nanA9GLENN GULLICKSON, TIMOTHYMANHATTAN AVENUE16.3386.833.01782000106920256260
92181184117NaN2021nanB9RAYMOND RECHTMANHATTAN AVENUE17.0070.003.0161100096660219300
................................................
3982531222530NaN20182C1M & N MANAGEMENT CORP168 SHERMAN AVENUE50.00150.005.0158400059850712800
39825412215356NaN20181B1PONDICHI CONSTANTIN10 VAN CORLEAR PLACE28.00133.002.06550002367434006
398264122029NaN20182D4420 WEST 206 STREET O416 WEST 206 STREET200.0099.006.060110001939502704950
3982651216444NaN20182C7175 REALTY ASSOCIATES4316 BROADWAY37.00103.005.0212200059400954900
3982831217563NaN20184K1PARTNERS 2004 LLC180 DYCKMAN STREET50.00100.001.023200002016001044000
\n", "

155480 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "Manhattandf" } }, "metadata": {}, "execution_count": 89 } ], "source": [ "Manhattandf" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "Ax3zLBLXgZrx" }, "outputs": [], "source": [ "# The remove_outliers_by_year_block was throwing an error saying year is both a index and a column\n", "# So here i am basically using .reset_index but it wasn't working so i am using this alt method\n", "\n", "Manhattandf = pd.DataFrame(Manhattandf.to_dict(orient='list'))" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "kkNTw8RMV6CI" }, "outputs": [], "source": [ "def remove_outliers_by_year_block(dataframe):\n", " # Group the DataFrame by 'year' and 'block'\n", " grouped = dataframe.groupby(['year', 'block'])\n", "\n", " # Function to remove outliers from each group\n", " def remove_outliers(group):\n", " q1 = group['fullval'].quantile(0.25)\n", " q3 = group['fullval'].quantile(0.75)\n", " iqr = q3 - q1\n", " lower_bound = q1 - 1.5 * iqr\n", " upper_bound = q3 + 1.5 * iqr\n", " return group[(group['fullval'] >= lower_bound) & (group['fullval'] <= upper_bound)]\n", "\n", " # Apply the function to each group and reset the index\n", " return grouped.apply(remove_outliers)\n", "\n", "# Apply the outlier removal function to the entire Manhattan df\n", "Manhattandf = remove_outliers_by_year_block(Manhattandf)\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "Dus8mbuMUu-o", "colab": { "base_uri": "https://localhost:8080/", "height": 455 }, "outputId": "93ce14b8-dfa9-4d21-c125-e3d1d13c2ad0" }, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " boro block lot easement year taxclass bldgcl \\\n", "year block \n", "2010 64 1162 1 64 106 NaN 2010 4 V1 \n", " 73 1163 1 73 14 NaN 2010 4 Y7 \n", " 120 1164 1 120 9 NaN 2010 3 U1 \n", " 142 1165 1 142 35 NaN 2010 4 W1 \n", " 1166 1 142 30 NaN 2010 4 Q1 \n", "... ... ... ... ... ... ... ... \n", "2021 1912 361 1 1912 1105 NaN 2021 nan R1 \n", " 1914 362 1 1914 40 NaN 2021 nan V1 \n", " 363 1 1914 41 NaN 2021 nan C0 \n", " 364 1 1914 55 NaN 2021 nan C5 \n", " 365 1 1914 56 NaN 2021 nan C5 \n", "\n", " owner staddr ltfront ltdepth \\\n", "year block \n", "2010 64 1162 NEW YORK CITY TRANSIT LIBERTY PLACE 3.00 19.00 \n", " 73 1163 DEPARTMENT OF BUSINES SOUTH STREET 166.00 140.00 \n", " 120 1164 DEPT OF TRANSPORTATIO FRANKFORT STREET 216.00 114.00 \n", " 142 1165 NaN GREENWICH STREET 0.00 0.00 \n", " 1166 DORMITORY AUTHORITYNY GREENWICH STREET 0.00 0.00 \n", "... ... ... ... ... \n", "2021 1912 361 BASDEVANT , JEROME LENOX AVENUE 0.00 0.00 \n", " 1914 362 DCAS WEST 130 STREET 16.67 99.92 \n", " 363 RASHID, FAZEELA ABDUL WEST 130 STREET 16.67 99.92 \n", " 364 STEPHENSON, DEBORAH A WEST 130 STREET 18.75 99.92 \n", " 365 160 W 130 REALTY LLC WEST 130 STREET 18.75 99.92 \n", "\n", " stories fullval avland avtot \n", "year block \n", "2010 64 1162 NaN 21800 9810 9810 \n", " 73 1163 2.0 3920000 1210500 1764000 \n", " 120 1164 NaN 1400000 630000 630000 \n", " 142 1165 NaN 9488893 1800002 4270002 \n", " 1166 NaN 25000000 11250000 11250000 \n", "... ... ... ... ... \n", "2021 1912 361 5.0 2962 1333 85751 \n", " 1914 362 0.0 672000 302400 302400 \n", " 363 3.0 646000 38760 118440 \n", " 364 3.0 35000 15750 195750 \n", " 365 3.0 213000 95850 241380 \n", "\n", "[139607 rows x 15 columns]" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
boroblockloteasementyeartaxclassbldgclownerstaddrltfrontltdepthstoriesfullvalavlandavtot
yearblock
2010641162164106NaN20104V1NEW YORK CITY TRANSITLIBERTY PLACE3.0019.00NaN2180098109810
73116317314NaN20104Y7DEPARTMENT OF BUSINESSOUTH STREET166.00140.002.0392000012105001764000
120116411209NaN20103U1DEPT OF TRANSPORTATIOFRANKFORT STREET216.00114.00NaN1400000630000630000
1421165114235NaN20104W1NaNGREENWICH STREET0.000.00NaN948889318000024270002
1166114230NaN20104Q1DORMITORY AUTHORITYNYGREENWICH STREET0.000.00NaN250000001125000011250000
......................................................
20211912361119121105NaN2021nanR1BASDEVANT , JEROMELENOX AVENUE0.000.005.02962133385751
19143621191440NaN2021nanV1DCASWEST 130 STREET16.6799.920.0672000302400302400
3631191441NaN2021nanC0RASHID, FAZEELA ABDULWEST 130 STREET16.6799.923.064600038760118440
3641191455NaN2021nanC5STEPHENSON, DEBORAH AWEST 130 STREET18.7599.923.03500015750195750
3651191456NaN2021nanC5160 W 130 REALTY LLCWEST 130 STREET18.7599.923.021300095850241380
\n", "

139607 rows × 15 columns

\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "Manhattandf" } }, "metadata": {}, "execution_count": 92 } ], "source": [ "Manhattandf" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 599 }, "id": "HJZMrga93T6N", "outputId": "6fa46dde-ffa3-4801-8a78-53a241c1b7f6" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "481576.25\n", "369509\n" ] }, { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": {} } ], "source": [ "# Used this to create a bell curve and discovered that there are outliers in price that are messing with certain predictions resulting in negative values\n", "# Negative values originate from the fact that there is data supporting a very low unrealistic price for a given year in a block that messes the prediction for other buildings in the same block\n", "# This was used for testing, this has no effect on the Manhattandf and the df referred here isnt used later\n", "# Information gathered here was used for other code\n", "\n", "import pandas as pd\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "from scipy.stats import norm\n", "\n", "df = Manhattandf\n", "\n", "block_input = 16\n", "year_input = 2017\n", "\n", "# Filter the DataFrame for the specified block and year\n", "filtered_df = df[(df['block'] == block_input) & (df['year'] == year_input)]\n", "\n", "\n", "prices = filtered_df['fullval'].dropna() # Drop NaN values\n", "\n", "mean = prices.mean()\n", "std = prices.std()\n", "# min = prices.min()\n", "\n", "# Generate a range of price values for the bell curve\n", "price_range = np.linspace(prices.min(), prices.max(), 100)\n", "\n", "bell_curve = norm.pdf(price_range, mean, std)\n", "\n", "print(mean)\n", "print(prices.min())\n", "\n", "plt.figure(figsize=(10, 6))\n", "plt.plot(price_range, bell_curve, color='blue')\n", "plt.title(f'Price Distribution of Buildings in Block {block_input} for the Year {year_input}')\n", "plt.xlabel('Price')\n", "plt.ylabel('Density')\n", "plt.grid(True)\n", "plt.show()\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "qeRHRR-jfu-M", "colab": { "base_uri": "https://localhost:8080/" }, "outputId": "b06a43cf-d526-4a29-c383-d62007c5b6f4" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Mean Squared Error: 69416622146159.26\n", "R-squared: 0.7917212784790276\n" ] } ], "source": [ "from sklearn.compose import ColumnTransformer\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.metrics import mean_squared_error, r2_score\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.ensemble import RandomForestRegressor\n", "from xgboost import XGBRegressor\n", "\n", "\n", "# Feature Selection and Engineering\n", "df_combined = Manhattandf\n", "\n", "# Removed staddr as we already have boro and block numbers so being further specific will serve little purpose\n", "# features_to_use = ['boro', 'block', 'lot', 'taxclass', 'bldgcl', 'staddr', 'ltfront', 'ltdepth', 'stories']\n", "features_to_use = ['block', 'taxclass', 'bldgcl', 'ltfront', 'ltdepth', 'stories', 'year']\n", "target = 'fullval'\n", "\n", "# Data Cleaning\n", "df_combined.dropna(subset=[target], inplace=True)\n", "\n", "# Convert categorical features to string\n", "# categorical_features = ['boro', 'taxclass', 'bldgcl', 'staddr'] # Assuming 'staddr' is categorical\n", "categorical_features = ['taxclass', 'bldgcl']\n", "for col in categorical_features:\n", " df_combined[col] = df_combined[col].astype(str)\n", "\n", "# Categorize features\n", "numerical_features = [f for f in features_to_use if f not in categorical_features]\n", "\n", "# Preprocessor for handling both categorical and numerical features\n", "preprocessor = ColumnTransformer(\n", " transformers=[\n", " ('num', Pipeline([\n", " ('imputer', SimpleImputer(strategy='mean')), # or 'median'\n", " ('scaler', StandardScaler())\n", " ]), numerical_features),\n", " ('cat', Pipeline([\n", " ('imputer', SimpleImputer(strategy='constant', fill_value='missing')),\n", " ('encoder', OneHotEncoder(handle_unknown='ignore'))\n", " ]), categorical_features)\n", " ]\n", ")\n", "'''\n", "# Linear Regression Model\n", "# Pipeline with model\n", "# Yielded .44 with 120,000 Values\n", "# Yielded .22 with 240,000 Values\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', LinearRegression())\n", "])\n", "'''\n", "'''\n", "# This model takes too long to load\n", "# Random Forest Regressor Model\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', RandomForestRegressor(n_estimators=100, random_state=42))\n", "])\n", "'''\n", "\n", "\n", "pipeline = Pipeline([\n", " ('preprocessor', preprocessor),\n", " ('model', XGBRegressor(objective='reg:squarederror', n_estimators=100, learning_rate=0.3, random_state=42))\n", "])\n", "\n", "\n", "# Model preparation\n", "X = df_combined[features_to_use]\n", "y = df_combined[target]\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n", "\n", "# Train the model\n", "pipeline.fit(X_train, y_train)\n", "\n", "# Predictions and Evaluation\n", "y_pred = pipeline.predict(X_test)\n", "mse = mean_squared_error(y_test, y_pred)\n", "r2 = r2_score(y_test, y_pred)\n", "\n", "print(f'Mean Squared Error: {mse}')\n", "print(f'R-squared: {r2}')\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 356 }, "id": "k8zidAEF0F6Y", "outputId": "faed793f-aae1-4aa9-a487-439dca5d3893" }, "outputs": [ { "output_type": "error", "ename": "KeyboardInterrupt", "evalue": "Interrupted by user", "traceback": [ "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", "\u001b[0;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)", "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 8\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0minput_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0muser_input\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m \u001b[0mblock\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter the block number (or N/A if unknown): \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mint\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 11\u001b[0m \u001b[0mtaxclass\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mget_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter the tax class (or N/A if unknown): \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 12\u001b[0m \u001b[0mbldgcl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter the building class (or N/A if unknown): \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m\u001b[0m in \u001b[0;36mget_input\u001b[0;34m(prompt, input_type)\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0;31m# Function to handle \"N/A\" input and convert it to None (which pandas interprets as np.nan)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 4\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprompt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_type\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0muser_input\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprompt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 6\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0muser_input\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'n/a'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 7\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36mraw_input\u001b[0;34m(self, prompt)\u001b[0m\n\u001b[1;32m 849\u001b[0m \u001b[0;34m\"raw_input was called, but this frontend does not support input requests.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 850\u001b[0m )\n\u001b[0;32m--> 851\u001b[0;31m return self._input_request(str(prompt),\n\u001b[0m\u001b[1;32m 852\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_ident\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 853\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_header\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36m_input_request\u001b[0;34m(self, prompt, ident, parent, password)\u001b[0m\n\u001b[1;32m 893\u001b[0m \u001b[0;32mexcept\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 894\u001b[0m \u001b[0;31m# re-raise KeyboardInterrupt, to truncate traceback\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 895\u001b[0;31m \u001b[0;32mraise\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Interrupted by user\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 896\u001b[0m \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 897\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwarning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Invalid Message:\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexc_info\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;31mKeyboardInterrupt\u001b[0m: Interrupted by user" ] } ], "source": [ "# Ask user for input and output cost prediction\n", "\n", "# Function to handle \"N/A\" input and convert it to None (which pandas interprets as np.nan)\n", "def get_input(prompt, input_type=str):\n", " user_input = input(prompt)\n", " if user_input.lower() == 'n/a':\n", " return None\n", " return input_type(user_input)\n", "\n", "block = get_input(\"Enter the block number (or N/A if unknown): \", int)\n", "taxclass =get_input(\"Enter the tax class (or N/A if unknown): \")\n", "bldgcl = get_input(\"Enter the building class (or N/A if unknown): \")\n", "ltfront = get_input(\"Enter the lot frontage (or N/A if unknown): \", float)\n", "ltdepth = get_input(\"Enter the lot depth (or N/A if unknown): \", float)\n", "stories = get_input(\"Enter the number of stories (or N/A if unknown): \", int)\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "410DuVfbw7jy" }, "outputs": [], "source": [ "'''\n", "year = 2018\n", "street_name = \"1 RIVER TERRACE\"\n", "block = 16\n", "lot_num = 3859\n", "fullval = 354180\n", "taxclass = 2\n", "building class = R4\n", "stores = 31\n", "'''" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "2dqfigtDbkj0" }, "outputs": [], "source": [ "year = get_input(\"Enter the year of purchase (or N/A if unknown): \", int)\n", "\n", "user_input = pd.DataFrame({\n", " 'block': [block],\n", " 'taxclass': [taxclass],\n", " 'bldgcl': [bldgcl],\n", " 'ltfront': [ltfront],\n", " 'ltdepth': [ltdepth],\n", " 'stories': [stories],\n", " 'year': [year]\n", "})\n", "\n", "# Check if user_input contains None, replace it with np.nan\n", "user_input = user_input.where(pd.notnull(user_input), None)\n", "\n", "predicted_value = pipeline.predict(user_input)\n", "\n", "print(f\"The predicted value of the building is: ${predicted_value[0]:,.2f}\" if predicted_value[0] is not None else \"Insufficient data to make a prediction.\")\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "i5QhKL68XYRZ" }, "outputs": [], "source": [ "print(predicted_value)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "X0V5dFIZOzMS" }, "outputs": [], "source": [ "print(\"To provide more accurate predictions and graphs, please provide the following extraneous metrics\")\n", "lot_num = get_input(\"Enter the lot number (or N/A if unknown): \", int)\n", "street_name = get_input(\"Enter the street_name (or N/A if unknown): \")\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "dyvJl278MRUj" }, "outputs": [], "source": [ "# Create empty dict of desired years use for processing\n", "\n", "cost_against_year = {}\n", "\n", "# Store ranges of years to implement in graphical displays\n", "for yeardecrement in range(8, 1, -1):\n", " cost_against_year[year - yeardecrement] = 0\n", "\n", "cost_against_year[year] = 0\n", "\n", "for yearincrement in range(1, 5):\n", " cost_against_year[year + yearincrement] = 0\n", "\n", "\n", "\n", "print(cost_against_year)\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "WDaIHmSfQewm" }, "outputs": [], "source": [ "# Loop through df to look for any existing values to base graphs off of rather than generating estimates using model\n", "# Then populate the rest of the years with predictions\n", "'''\n", "existing_entries_year = []\n", "\n", "for index, row in df4.iterrows():\n", " if row['year'] in cost_against_year and row['staddr'] == street_name and row['lot'] == lot_num and row['block'] == block:\n", " cost_against_year[row['year']] += row['fullval']\n", " existing_entries_year.append(row['year'])\n", "'''" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "PqasiPrVa5zg" }, "outputs": [], "source": [ "for key, value in cost_against_year.items():\n", " print(value)\n", " if value == 0:\n", " prediction_year_df = pd.DataFrame({\n", " 'block': [block],\n", " 'taxclass': [taxclass],\n", " 'bldgcl': [bldgcl],\n", " 'ltfront': [ltfront],\n", " 'ltdepth': [ltdepth],\n", " 'stories': [stories],\n", " 'year': [key]\n", " })\n", " prediction_year_df.where(pd.notnull(user_input), None)\n", "\n", " predicition_year = pipeline.predict(prediction_year_df)\n", " print(key, \"year\")\n", " print(predicition_year)\n", " cost_against_year[key] = predicition_year[0]" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "W5iqplJOWFHF" }, "outputs": [], "source": [ "print(cost_against_year)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "sEJjjqjKzXOx" }, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "1pDk5-y87T-W" }, "outputs": [], "source": [ "# Heatmap of high profit and low profit areas red adn blue\n", "# Genrerate change of cost over the years\n", "# Line graphs of price change\n", "# Use gpt to update with real time events and produce a better humanlike itnerface" ] } ], "metadata": { "colab": { "provenance": [] }, "kernelspec": { "display_name": "Python 3", "name": "python3" }, "language_info": { "name": "python" } }, "nbformat": 4, "nbformat_minor": 0 }