{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# 🚲 Dublin Bikes Availability Prediction Model\n", "\n", "## Project Objective\n", "Train a machine learning model using historical bike and weather data \n", "to predict the **number of available bikes** at a station for a specific time.\n", "\n", "## Dataset\n", "- File name: `final_merged_data.csv`\n", "- Content: merged bike station history data + weather data\n", "\n", "## Notebook Structure\n", "1. Environment setup & data loading\n", "2. Data exploration (EDA)\n", "3. Data cleaning & preprocessing\n", "4. Feature engineering\n", "5. Model training & comparison\n", "6. Model evaluation\n", "7. Save the best model" ], "id": "1340a012" }, { "cell_type": "code", "metadata": {}, "source": [ "# Data processing\n", "import pandas as pd\n", "import numpy as np\n", "\n", "# Visualization\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "from matplotlib import font_manager\n", "\n", "# Machine learning\n", "from sklearn.linear_model import LinearRegression, LogisticRegression\n", "from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor\n", "from sklearn.tree import DecisionTreeRegressor\n", "from sklearn.model_selection import train_test_split, cross_val_score\n", "from sklearn.metrics import mean_absolute_error, r2_score, mean_squared_error\n", "from sklearn.preprocessing import StandardScaler\n", "from sklearn.pipeline import Pipeline\n", "\n", "# Model persistence\n", "import pickle\n", "import joblib\n", "\n", "# Other\n", "import warnings\n", "warnings.filterwarnings('ignore')\n", "\n", "# Set visualization style\n", "plt.style.use('seaborn-v0_8')\n", "sns.set_palette(\"husl\")\n", "\n", "# Chinese font configuration\n", "plt.rcParams['font.sans-serif'] = ['Arial Unicode MS']\n", "plt.rcParams['axes.unicode_minus'] = False\n", "\n", "# Set pandas display options\n", "pd.set_option('display.max_columns', None)\n", "pd.set_option('display.float_format', lambda x: '%.3f' % x)\n", "\n", "print(\"✅ All libraries imported successfully\")" ], "execution_count": 39, "outputs": [ { "output_type": "stream", "text": [ "✅ All libraries imported successfully\n" ] } ], "id": "865007da" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 1. Data Loading\n", "\n", "Read `final_merged_data.csv` and perform an initial inspection:\n", "- Number of rows and columns\n", "- Preview of the first few rows\n", "- Column names and data types\n", "- Basic statistics" ], "id": "79708e38" }, { "cell_type": "code", "metadata": {}, "source": [ "# Load data\n", "df = pd.read_csv('final_merged_data.csv')\n", "\n", "# Basic information\n", "print(f\"📊 Dataset size: {df.shape[0]} rows x {df.shape[1]} columns\")\n", "print(f\"\\n📋 Column name list:\")\n", "print(df.columns.tolist())\n", "print(f\"\\n🔍 Data types:\")\n", "print(df.dtypes)" ], "execution_count": 40, "outputs": [ { "output_type": "stream", "text": [ "📊 Dataset size: 298946 rows x 78 columns\n", "\n", "📋 Column name list:\n", "['last_reported', 'station_id', 'num_bikes_available', 'num_docks_available', 'is_installed', 'is_renting', 'is_returning', 'name', 'address', 'lat', 'lon', 'capacity', 'stno', 'year', 'month', 'day', 'hour', 'minute', 'max_air_temp_quality_indicator', 'max_air_temperature_celsius', 'min_air_temp_quality_indicator', 'min_air_temperature_celsius', 'air_temp_std_quality_indicator', 'air_temperature_std_deviation', 'max_grass_temp_quality_indicator', 'max_grass_temperature_celsius', 'min_grass_temp_quality_indicator', 'min_grass_temperature_celsius', 'grass_temp_std_quality_indicator', 'grass_temperature_std_deviation', 'max_soil_temp_5cm_quality_indicator', 'max_soil_temperature_5cm_celsius', 'min_soil_temp_5cm_quality_indicator', 'min_soil_temperature_5cm_celsius', 'soil_temp_std_5cm_quality_indicator', 'soil_temperature_std_deviation_5cm', 'max_soil_temp_10cm_quality_indicator', 'max_soil_temperature_10cm_celsius', 'min_soil_temp_10cm_quality_indicator', 'min_soil_temperature_10cm_celsius', 'soil_temp_std_10cm_quality_indicator', 'soil_temperature_std_deviation_10cm', 'max_soil_temp_20cm_quality_indicator', 'max_soil_temperature_20cm_celsius', 'min_soil_temp_20cm_quality_indicator', 'min_soil_temperature_20cm_celsius', 'soil_temp_std_20cm_quality_indicator', 'soil_temperature_std_deviation_20cm', 'max_earth_temp_30cm_quality_indicator', 'max_earth_temperature_30cm_celsius', 'min_earth_temp_30cm_quality_indicator', 'min_earth_temperature_30cm_celsius', 'earth_temp_std_30cm_quality_indicator', 'earth_temperature_std_deviation_30cm', 'max_earth_temp_50cm_quality_indicator', 'max_earth_temperature_50cm_celsius', 'min_earth_temp_50cm_quality_indicator', 'min_earth_temperature_50cm_celsius', 'earth_temp_std_50cm_quality_indicator', 'earth_temperature_std_deviation_50cm', 'max_earth_temp_100cm_quality_indicator', 'max_earth_temperature_100cm_celsius', 'min_earth_temp_100cm_quality_indicator', 'min_earth_temperature_100cm_celsius', 'earth_temp_std_100cm_quality_indicator', 'earth_temperature_std_deviation_100cm', 'max_humidity_quality_indicator', 'max_relative_humidity_percent', 'min_humidity_quality_indicator', 'min_relative_humidity_percent', 'humidity_std_quality_indicator', 'relative_humidity_std_deviation', 'max_pressure_quality_indicator', 'max_barometric_pressure_hpa', 'min_pressure_quality_indicator', 'min_barometric_pressure_hpa', 'pressure_std_quality_indicator', 'barometric_pressure_std_deviation']\n", "\n", "🔍 Data types:\n", "last_reported str\n", "station_id int64\n", "num_bikes_available int64\n", "num_docks_available int64\n", "is_installed bool\n", " ... \n", "max_barometric_pressure_hpa float64\n", "min_pressure_quality_indicator int64\n", "min_barometric_pressure_hpa float64\n", "pressure_std_quality_indicator int64\n", "barometric_pressure_std_deviation float64\n", "Length: 78, dtype: object\n" ] } ], "id": "c2a3bc97" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 2. Initial Data Exploration (EDA)\n", "\n", "First, inspect the basic status of the data:\n", "- Missing value distribution\n", "- Target variable distribution\n", "- Preview the first few rows" ], "id": "4c90e796" }, { "cell_type": "code", "metadata": {}, "source": [ "# Preview first 5 rows\n", "print(\"📋 Preview of first 5 rows:\")\n", "df.head()" ], "execution_count": 41, "outputs": [ { "output_type": "stream", "text": [ "📋 Preview of first 5 rows:\n" ] }, { "output_type": "execute_result", "data": { "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
last_reportedstation_idnum_bikes_availablenum_docks_availableis_installedis_rentingis_returningnameaddresslatloncapacitystnoyearmonthdayhourminutemax_air_temp_quality_indicatormax_air_temperature_celsiusmin_air_temp_quality_indicatormin_air_temperature_celsiusair_temp_std_quality_indicatorair_temperature_std_deviationmax_grass_temp_quality_indicatormax_grass_temperature_celsiusmin_grass_temp_quality_indicatormin_grass_temperature_celsiusgrass_temp_std_quality_indicatorgrass_temperature_std_deviationmax_soil_temp_5cm_quality_indicatormax_soil_temperature_5cm_celsiusmin_soil_temp_5cm_quality_indicatormin_soil_temperature_5cm_celsiussoil_temp_std_5cm_quality_indicatorsoil_temperature_std_deviation_5cmmax_soil_temp_10cm_quality_indicatormax_soil_temperature_10cm_celsiusmin_soil_temp_10cm_quality_indicatormin_soil_temperature_10cm_celsiussoil_temp_std_10cm_quality_indicatorsoil_temperature_std_deviation_10cmmax_soil_temp_20cm_quality_indicatormax_soil_temperature_20cm_celsiusmin_soil_temp_20cm_quality_indicatormin_soil_temperature_20cm_celsiussoil_temp_std_20cm_quality_indicatorsoil_temperature_std_deviation_20cmmax_earth_temp_30cm_quality_indicatormax_earth_temperature_30cm_celsiusmin_earth_temp_30cm_quality_indicatormin_earth_temperature_30cm_celsiusearth_temp_std_30cm_quality_indicatorearth_temperature_std_deviation_30cmmax_earth_temp_50cm_quality_indicatormax_earth_temperature_50cm_celsiusmin_earth_temp_50cm_quality_indicatormin_earth_temperature_50cm_celsiusearth_temp_std_50cm_quality_indicatorearth_temperature_std_deviation_50cmmax_earth_temp_100cm_quality_indicatormax_earth_temperature_100cm_celsiusmin_earth_temp_100cm_quality_indicatormin_earth_temperature_100cm_celsiusearth_temp_std_100cm_quality_indicatorearth_temperature_std_deviation_100cmmax_humidity_quality_indicatormax_relative_humidity_percentmin_humidity_quality_indicatormin_relative_humidity_percenthumidity_std_quality_indicatorrelative_humidity_std_deviationmax_pressure_quality_indicatormax_barometric_pressure_hpamin_pressure_quality_indicatormin_barometric_pressure_hpapressure_std_quality_indicatorbarometric_pressure_std_deviation
02024-12-01 00:10:0010151TrueTrueTrueDAME STREETDame Street53.344-6.267161752024121010014.010013.90000.033011.930011.73000.056010.150010.11000.00809.58009.56000.00508.83008.80000.00708.88008.85000.00808.50008.47000.00809.87009.83000.008084.300083.20000.28401002.56001002.26000.083
12024-12-01 00:10:00100178TrueTrueTrueHEUSTON BRIDGE (SOUTH)Heuston Bridge (South)53.347-6.292251752024121010014.010013.90000.033011.930011.73000.056010.150010.11000.00809.58009.56000.00508.83008.80000.00708.88008.85000.00808.50008.47000.00809.87009.83000.008084.300083.20000.28401002.56001002.26000.083
22024-12-01 00:10:00109209TrueTrueTrueBUCKINGHAM STREET LOWERBuckingham Street Lower53.353-6.249291752024121010014.010013.90000.033011.930011.73000.056010.150010.11000.00809.58009.56000.00508.83008.80000.00708.88008.85000.00808.50008.47000.00809.87009.83000.008084.300083.20000.28401002.56001002.26000.083
32024-12-01 00:10:0011129TrueTrueTrueEARLSFORT TERRACEEarlsfort Terrace53.334-6.259301752024121010014.010013.90000.033011.930011.73000.056010.150010.11000.00809.58009.56000.00508.83008.80000.00708.88008.85000.00808.50008.47000.00809.87009.83000.008084.300083.20000.28401002.56001002.26000.083
42024-12-01 00:10:00114436TrueTrueTrueWILTON TERRACE (PARK)Wilton Terrace (Park)53.334-6.248401752024121010014.010013.90000.033011.930011.73000.056010.150010.11000.00809.58009.56000.00508.83008.80000.00708.88008.85000.00808.50008.47000.00809.87009.83000.008084.300083.20000.28401002.56001002.26000.083
\n", "
" ], "text/plain": [ " last_reported station_id num_bikes_available num_docks_available \\\n", "0 2024-12-01 00:10:00 10 15 1 \n", "1 2024-12-01 00:10:00 100 17 8 \n", "2 2024-12-01 00:10:00 109 20 9 \n", "3 2024-12-01 00:10:00 11 1 29 \n", "4 2024-12-01 00:10:00 114 4 36 \n", "\n", " is_installed is_renting is_returning name \\\n", "0 True True True DAME STREET \n", "1 True True True HEUSTON BRIDGE (SOUTH) \n", "2 True True True BUCKINGHAM STREET LOWER \n", "3 True True True EARLSFORT TERRACE \n", "4 True True True WILTON TERRACE (PARK) \n", "\n", " address lat lon capacity stno year month day \\\n", "0 Dame Street 53.344 -6.267 16 175 2024 12 1 \n", "1 Heuston Bridge (South) 53.347 -6.292 25 175 2024 12 1 \n", "2 Buckingham Street Lower 53.353 -6.249 29 175 2024 12 1 \n", "3 Earlsfort Terrace 53.334 -6.259 30 175 2024 12 1 \n", "4 Wilton Terrace (Park) 53.334 -6.248 40 175 2024 12 1 \n", "\n", " hour minute max_air_temp_quality_indicator max_air_temperature_celsius \\\n", "0 0 10 0 14.010 \n", "1 0 10 0 14.010 \n", "2 0 10 0 14.010 \n", "3 0 10 0 14.010 \n", "4 0 10 0 14.010 \n", "\n", " min_air_temp_quality_indicator min_air_temperature_celsius \\\n", "0 0 13.900 \n", "1 0 13.900 \n", "2 0 13.900 \n", "3 0 13.900 \n", "4 0 13.900 \n", "\n", " air_temp_std_quality_indicator air_temperature_std_deviation \\\n", "0 0 0.033 \n", "1 0 0.033 \n", "2 0 0.033 \n", "3 0 0.033 \n", "4 0 0.033 \n", "\n", " max_grass_temp_quality_indicator max_grass_temperature_celsius \\\n", "0 0 11.930 \n", "1 0 11.930 \n", "2 0 11.930 \n", "3 0 11.930 \n", "4 0 11.930 \n", "\n", " min_grass_temp_quality_indicator min_grass_temperature_celsius \\\n", "0 0 11.730 \n", "1 0 11.730 \n", "2 0 11.730 \n", "3 0 11.730 \n", "4 0 11.730 \n", "\n", " grass_temp_std_quality_indicator grass_temperature_std_deviation \\\n", "0 0 0.056 \n", "1 0 0.056 \n", "2 0 0.056 \n", "3 0 0.056 \n", "4 0 0.056 \n", "\n", " max_soil_temp_5cm_quality_indicator max_soil_temperature_5cm_celsius \\\n", "0 0 10.150 \n", "1 0 10.150 \n", "2 0 10.150 \n", "3 0 10.150 \n", "4 0 10.150 \n", "\n", " min_soil_temp_5cm_quality_indicator min_soil_temperature_5cm_celsius \\\n", "0 0 10.110 \n", "1 0 10.110 \n", "2 0 10.110 \n", "3 0 10.110 \n", "4 0 10.110 \n", "\n", " soil_temp_std_5cm_quality_indicator soil_temperature_std_deviation_5cm \\\n", "0 0 0.008 \n", "1 0 0.008 \n", "2 0 0.008 \n", "3 0 0.008 \n", "4 0 0.008 \n", "\n", " max_soil_temp_10cm_quality_indicator max_soil_temperature_10cm_celsius \\\n", "0 0 9.580 \n", "1 0 9.580 \n", "2 0 9.580 \n", "3 0 9.580 \n", "4 0 9.580 \n", "\n", " min_soil_temp_10cm_quality_indicator min_soil_temperature_10cm_celsius \\\n", "0 0 9.560 \n", "1 0 9.560 \n", "2 0 9.560 \n", "3 0 9.560 \n", "4 0 9.560 \n", "\n", " soil_temp_std_10cm_quality_indicator soil_temperature_std_deviation_10cm \\\n", "0 0 0.005 \n", "1 0 0.005 \n", "2 0 0.005 \n", "3 0 0.005 \n", "4 0 0.005 \n", "\n", " max_soil_temp_20cm_quality_indicator max_soil_temperature_20cm_celsius \\\n", "0 0 8.830 \n", "1 0 8.830 \n", "2 0 8.830 \n", "3 0 8.830 \n", "4 0 8.830 \n", "\n", " min_soil_temp_20cm_quality_indicator min_soil_temperature_20cm_celsius \\\n", "0 0 8.800 \n", "1 0 8.800 \n", "2 0 8.800 \n", "3 0 8.800 \n", "4 0 8.800 \n", "\n", " soil_temp_std_20cm_quality_indicator soil_temperature_std_deviation_20cm \\\n", "0 0 0.007 \n", "1 0 0.007 \n", "2 0 0.007 \n", "3 0 0.007 \n", "4 0 0.007 \n", "\n", " max_earth_temp_30cm_quality_indicator max_earth_temperature_30cm_celsius \\\n", "0 0 8.880 \n", "1 0 8.880 \n", "2 0 8.880 \n", "3 0 8.880 \n", "4 0 8.880 \n", "\n", " min_earth_temp_30cm_quality_indicator min_earth_temperature_30cm_celsius \\\n", "0 0 8.850 \n", "1 0 8.850 \n", "2 0 8.850 \n", "3 0 8.850 \n", "4 0 8.850 \n", "\n", " earth_temp_std_30cm_quality_indicator \\\n", "0 0 \n", "1 0 \n", "2 0 \n", "3 0 \n", "4 0 \n", "\n", " earth_temperature_std_deviation_30cm \\\n", "0 0.008 \n", "1 0.008 \n", "2 0.008 \n", "3 0.008 \n", "4 0.008 \n", "\n", " max_earth_temp_50cm_quality_indicator max_earth_temperature_50cm_celsius \\\n", "0 0 8.500 \n", "1 0 8.500 \n", "2 0 8.500 \n", "3 0 8.500 \n", "4 0 8.500 \n", "\n", " min_earth_temp_50cm_quality_indicator min_earth_temperature_50cm_celsius \\\n", "0 0 8.470 \n", "1 0 8.470 \n", "2 0 8.470 \n", "3 0 8.470 \n", "4 0 8.470 \n", "\n", " earth_temp_std_50cm_quality_indicator \\\n", "0 0 \n", "1 0 \n", "2 0 \n", "3 0 \n", "4 0 \n", "\n", " earth_temperature_std_deviation_50cm \\\n", "0 0.008 \n", "1 0.008 \n", "2 0.008 \n", "3 0.008 \n", "4 0.008 \n", "\n", " max_earth_temp_100cm_quality_indicator \\\n", "0 0 \n", "1 0 \n", "2 0 \n", "3 0 \n", "4 0 \n", "\n", " max_earth_temperature_100cm_celsius \\\n", "0 9.870 \n", "1 9.870 \n", "2 9.870 \n", "3 9.870 \n", "4 9.870 \n", "\n", " min_earth_temp_100cm_quality_indicator \\\n", "0 0 \n", "1 0 \n", "2 0 \n", "3 0 \n", "4 0 \n", "\n", " min_earth_temperature_100cm_celsius \\\n", "0 9.830 \n", "1 9.830 \n", "2 9.830 \n", "3 9.830 \n", "4 9.830 \n", "\n", " earth_temp_std_100cm_quality_indicator \\\n", "0 0 \n", "1 0 \n", "2 0 \n", "3 0 \n", "4 0 \n", "\n", " earth_temperature_std_deviation_100cm max_humidity_quality_indicator \\\n", "0 0.008 0 \n", "1 0.008 0 \n", "2 0.008 0 \n", "3 0.008 0 \n", "4 0.008 0 \n", "\n", " max_relative_humidity_percent min_humidity_quality_indicator \\\n", "0 84.300 0 \n", "1 84.300 0 \n", "2 84.300 0 \n", "3 84.300 0 \n", "4 84.300 0 \n", "\n", " min_relative_humidity_percent humidity_std_quality_indicator \\\n", "0 83.200 0 \n", "1 83.200 0 \n", "2 83.200 0 \n", "3 83.200 0 \n", "4 83.200 0 \n", "\n", " relative_humidity_std_deviation max_pressure_quality_indicator \\\n", "0 0.284 0 \n", "1 0.284 0 \n", "2 0.284 0 \n", "3 0.284 0 \n", "4 0.284 0 \n", "\n", " max_barometric_pressure_hpa min_pressure_quality_indicator \\\n", "0 1002.560 0 \n", "1 1002.560 0 \n", "2 1002.560 0 \n", "3 1002.560 0 \n", "4 1002.560 0 \n", "\n", " min_barometric_pressure_hpa pressure_std_quality_indicator \\\n", "0 1002.260 0 \n", "1 1002.260 0 \n", "2 1002.260 0 \n", "3 1002.260 0 \n", "4 1002.260 0 \n", "\n", " barometric_pressure_std_deviation \n", "0 0.083 \n", "1 0.083 \n", "2 0.083 \n", "3 0.083 \n", "4 0.083 " ] } } ], "id": "242a3c70" }, { "cell_type": "code", "metadata": {}, "source": [ "# Missing value summary\n", "missing = df.isnull().sum()\n", "missing_pct = (missing / len(df) * 100).round(2)\n", "\n", "missing_df = pd.DataFrame({\n", " 'Missing Count': missing,\n", " 'Missing Ratio (%)': missing_pct\n", "}).query('`Missing Count` > 0').sort_values('Missing Ratio (%)', ascending=False)\n", "\n", "print(f\"⚠️ Total {len(missing_df)} columns with missing values:\\n\")\n", "print(missing_df)" ], "execution_count": 42, "outputs": [ { "output_type": "stream", "text": [ "⚠️ Total 0 columns with missing values:\n", "\n", "Empty DataFrame\n", "Columns: [Missing Count, Missing Ratio (%)]\n", "Index: []\n" ] } ], "id": "4afef0ad" }, { "cell_type": "code", "metadata": {}, "source": [ "# Target variable distribution\n", "print(\"📊 Target variable num_bikes_available statistics:\")\n", "print(df['num_bikes_available'].describe())\n", "\n", "plt.figure(figsize=(10, 4))\n", "plt.subplot(1, 2, 1)\n", "df['num_bikes_available'].hist(bins=30, color='steelblue', edgecolor='white')\n", "plt.title('Distribution of Available Bikes')\n", "plt.xlabel('Available Bikes')\n", "plt.ylabel('Frequency')\n", "\n", "plt.subplot(1, 2, 2)\n", "df['num_bikes_available'].plot(kind='box', color='steelblue')\n", "plt.title('Box Plot of Available Bikes')\n", "\n", "plt.tight_layout()\n", "plt.show()" ], "execution_count": 43, "outputs": [ { "output_type": "stream", "text": [ "📊 Target variable num_bikes_available statistics:\n", "count 298946.000\n", "mean 12.205\n", "std 9.762\n", "min 0.000\n", "25% 4.000\n", "50% 11.000\n", "75% 19.000\n", "max 40.000\n", "Name: num_bikes_available, dtype: float64\n" ] }, { "output_type": "display_data", "data": { "image/png": "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", "text/plain": [ "
" ] } } ], "id": "d051f9ba" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 3. Feature Selection\n", "\n", "Based on column analysis, keep the following meaningful features:\n", "\n", "**Time features:** `hour`, `month`, `day`, `year`\n", "\n", "**Station features:** `station_id`, `capacity`, `lat`, `lon`\n", "\n", "**Weather features:**\n", "- Temperature:`max_air_temperature_celsius`, `min_air_temperature_celsius`\n", "- Humidity:`max_relative_humidity_percent`, `min_relative_humidity_percent`\n", "- Pressure:`max_barometric_pressure_hpa`, `min_barometric_pressure_hpa`\n", "\n", "**Reasons for removal:**\n", "- All `quality_indicator` columns -> quality flags, not actual numeric values\n", "- All `std_deviation` columns -> redundant information\n", "- Soil/grass temperatures → low relevance to bike usage\n", "- `name`, `address`, `last_reported` → text columns, not used for modeling\n", "- `is_installed`, `is_renting`, `is_returning` → almost always True, low discriminative power\n", "\n", "**Target variable:** `num_bikes_available`" ], "id": "5919a790" }, { "cell_type": "code", "metadata": {}, "source": [ "# Select required columns\n", "selected_columns = [\n", " # Time features\n", " 'hour', 'month', 'day', 'year',\n", " \n", " # Station features\n", " 'station_id', 'capacity', 'lat', 'lon',\n", " \n", " # Weather features\n", " 'max_air_temperature_celsius', 'min_air_temperature_celsius',\n", " 'max_relative_humidity_percent', 'min_relative_humidity_percent',\n", " 'max_barometric_pressure_hpa', 'min_barometric_pressure_hpa',\n", " \n", " # Target variable\n", " 'num_bikes_available'\n", "]\n", "\n", "# Create a new DataFrame\n", "df_model = df[selected_columns].copy()\n", "\n", "print(f\"✅ Feature selection completed\")\n", "print(f\"📊 New dataset size: {df_model.shape[0]} rows x {df_model.shape[1]} columns\")\n", "print(f\"\\n📋 Data types of each column:\")\n", "print(df_model.dtypes)\n", "print(f\"\\n📋 First 5 rows preview:\")\n", "df_model.head()" ], "execution_count": 44, "outputs": [ { "output_type": "stream", "text": [ "✅ Feature selection completed\n", "📊 New dataset size: 298946 rows x 15 columns\n", "\n", "📋 Data types of each column:\n", "hour int64\n", "month int64\n", "day int64\n", "year int64\n", "station_id int64\n", "capacity int64\n", "lat float64\n", "lon float64\n", "max_air_temperature_celsius float64\n", "min_air_temperature_celsius float64\n", "max_relative_humidity_percent float64\n", "min_relative_humidity_percent float64\n", "max_barometric_pressure_hpa float64\n", "min_barometric_pressure_hpa float64\n", "num_bikes_available int64\n", "dtype: object\n", "\n", "📋 First 5 rows preview:\n" ] }, { "output_type": "execute_result", "data": { "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", "
hourmonthdayyearstation_idcapacitylatlonmax_air_temperature_celsiusmin_air_temperature_celsiusmax_relative_humidity_percentmin_relative_humidity_percentmax_barometric_pressure_hpamin_barometric_pressure_hpanum_bikes_available
001212024101653.344-6.26714.01013.90084.30083.2001002.5601002.26015
1012120241002553.347-6.29214.01013.90084.30083.2001002.5601002.26017
2012120241092953.353-6.24914.01013.90084.30083.2001002.5601002.26020
301212024113053.334-6.25914.01013.90084.30083.2001002.5601002.2601
4012120241144053.334-6.24814.01013.90084.30083.2001002.5601002.2604
\n", "
" ], "text/plain": [ " hour month day year station_id capacity lat lon \\\n", "0 0 12 1 2024 10 16 53.344 -6.267 \n", "1 0 12 1 2024 100 25 53.347 -6.292 \n", "2 0 12 1 2024 109 29 53.353 -6.249 \n", "3 0 12 1 2024 11 30 53.334 -6.259 \n", "4 0 12 1 2024 114 40 53.334 -6.248 \n", "\n", " max_air_temperature_celsius min_air_temperature_celsius \\\n", "0 14.010 13.900 \n", "1 14.010 13.900 \n", "2 14.010 13.900 \n", "3 14.010 13.900 \n", "4 14.010 13.900 \n", "\n", " max_relative_humidity_percent min_relative_humidity_percent \\\n", "0 84.300 83.200 \n", "1 84.300 83.200 \n", "2 84.300 83.200 \n", "3 84.300 83.200 \n", "4 84.300 83.200 \n", "\n", " max_barometric_pressure_hpa min_barometric_pressure_hpa \\\n", "0 1002.560 1002.260 \n", "1 1002.560 1002.260 \n", "2 1002.560 1002.260 \n", "3 1002.560 1002.260 \n", "4 1002.560 1002.260 \n", "\n", " num_bikes_available \n", "0 15 \n", "1 17 \n", "2 20 \n", "3 1 \n", "4 4 " ] } } ], "id": "1606bf53" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 4. Feature engineering\n", "\n", "Based on existing features, add the following derived features:\n", "\n", "- `day_of_week`: day of week (0=Monday, 6=Sunday), captures weekday/weekend patterns\n", "- `is_weekend`: whether it is weekend (0 or 1), a direct weekend flag\n", "- `avg_temperature`: average of max and min temperature as representative temperature\n", "- `avg_humidity`: average of max and min humidity\n", "- `avg_pressure`: average of max and min pressure" ], "id": "1db37db4" }, { "cell_type": "code", "metadata": {}, "source": [ "# Add derived features\n", "# Build datetime from year/month/day to extract day of week\n", "df_model['day_of_week'] = pd.to_datetime(df[['year', 'month', 'day']]).dt.dayofweek\n", "\n", "# Weekend flag\n", "df_model['is_weekend'] = df_model['day_of_week'].apply(lambda x: 1 if x >= 5 else 0)\n", "\n", "# Average temperature\n", "df_model['avg_temperature'] = (df_model['max_air_temperature_celsius'] + df_model['min_air_temperature_celsius']) / 2\n", "\n", "# Average humidity\n", "df_model['avg_humidity'] = (df_model['max_relative_humidity_percent'] + df_model['min_relative_humidity_percent']) / 2\n", "\n", "# Average pressure\n", "df_model['avg_pressure'] = (df_model['max_barometric_pressure_hpa'] + df_model['min_barometric_pressure_hpa']) / 2\n", "\n", "print(f\"✅ Feature engineering completed\")\n", "print(f\"📊 New dataset size: {df_model.shape[0]} rows x {df_model.shape[1]} columns\")\n", "print(f\"\\nSummary statistics of new features:\")\n", "print(df_model[['day_of_week', 'is_weekend', 'avg_temperature', 'avg_humidity', 'avg_pressure']].describe())" ], "execution_count": 45, "outputs": [ { "output_type": "stream", "text": [ "✅ Feature engineering completed\n", "📊 New dataset size: 298946 rows x 20 columns\n", "\n", "Summary statistics of new features:\n", " day_of_week is_weekend avg_temperature avg_humidity avg_pressure\n", "count 298946.000 298946.000 298946.000 298946.000 298946.000\n", "mean 2.923 0.282 7.786 84.285 1014.773\n", "std 2.061 0.450 3.129 8.005 11.859\n", "min 0.000 0.000 -3.484 55.070 975.035\n", "25% 1.000 0.000 5.676 79.770 1005.915\n", "50% 3.000 0.000 7.872 85.350 1017.125\n", "75% 5.000 1.000 10.125 89.600 1022.615\n", "max 6.000 1.000 14.630 98.850 1035.775\n" ] } ], "id": "904f6cb4" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 5. Correlation Analysis\n", "\n", "Examine correlations between each feature and the target variable `num_bikes_available`\n", "to identify which features are most helpful for prediction." ], "id": "57ffb8a8" }, { "cell_type": "code", "metadata": {}, "source": [ "# Compute correlation matrix\n", "corr_matrix = df_model.corr()\n", "\n", "# Extract and sort correlations with target variable\n", "target_corr = corr_matrix['num_bikes_available'].drop('num_bikes_available').sort_values(ascending=False)\n", "\n", "print(\"📊 Correlation of each feature with num_bikes_available:\")\n", "print(target_corr)\n", "\n", "# Visualization\n", "plt.figure(figsize=(10, 6))\n", "colors = ['steelblue' if x > 0 else 'tomato' for x in target_corr.values]\n", "target_corr.plot(kind='bar', color=colors)\n", "plt.title('Correlation of Features with Available Bikes')\n", "plt.xlabel('Feature')\n", "plt.ylabel('Correlation Coefficient')\n", "plt.xticks(rotation=45, ha='right')\n", "plt.axhline(y=0, color='black', linewidth=0.8)\n", "plt.tight_layout()\n", "plt.show()" ], "execution_count": 46, "outputs": [ { "output_type": "stream", "text": [ "📊 Correlation of each feature with num_bikes_available:\n", "capacity 0.205\n", "lon 0.125\n", "day_of_week 0.010\n", "min_relative_humidity_percent 0.009\n", "avg_humidity 0.009\n", "max_relative_humidity_percent 0.009\n", "is_weekend 0.007\n", "day 0.004\n", "max_air_temperature_celsius 0.002\n", "avg_temperature 0.002\n", "min_air_temperature_celsius 0.002\n", "station_id -0.001\n", "max_barometric_pressure_hpa -0.002\n", "avg_pressure -0.002\n", "min_barometric_pressure_hpa -0.002\n", "hour -0.005\n", "lat -0.138\n", "month NaN\n", "year NaN\n", "Name: num_bikes_available, dtype: float64\n" ] }, { "output_type": "display_data", "data": { "image/png": "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", "text/plain": [ "
" ] } } ], "id": "4f581f4c" }, { "cell_type": "code", "metadata": {}, "source": [ "# Check unique values of month and year\n", "print(\"month unique values:\", df_model['month'].unique())\n", "print(\"year unique values:\", df_model['year'].unique())\n", "\n", "# If either column has a single value, variance is 0 and correlation cannot be computed\n", "print(\"\\nmonth variance:\", df_model['month'].var())\n", "print(\"year variance:\", df_model['year'].var())" ], "execution_count": 47, "outputs": [ { "output_type": "stream", "text": [ "month unique values: [12]\n", "year unique values: [2024]\n", "\n", "month variance: 0.0\n", "year variance: 0.0\n" ] } ], "id": "2f5ba491" }, { "cell_type": "code", "metadata": {}, "source": [ "# Drop zero-variance columns (not useful for prediction)\n", "df_model = df_model.drop(columns=['month', 'year'])\n", "\n", "print(f\"✅ Removed month and year columns\")\n", "print(f\"📊 Current dataset size: {df_model.shape[0]} rows x {df_model.shape[1]} columns\")\n", "print(f\"\\n📋 Current feature list：\")\n", "print(df_model.columns.tolist())" ], "execution_count": 48, "outputs": [ { "output_type": "stream", "text": [ "✅ Removed month and year columns\n", "📊 Current dataset size: 298946 rows x 18 columns\n", "\n", "📋 Current feature list：\n", "['hour', 'day', 'station_id', 'capacity', 'lat', 'lon', 'max_air_temperature_celsius', 'min_air_temperature_celsius', 'max_relative_humidity_percent', 'min_relative_humidity_percent', 'max_barometric_pressure_hpa', 'min_barometric_pressure_hpa', 'num_bikes_available', 'day_of_week', 'is_weekend', 'avg_temperature', 'avg_humidity', 'avg_pressure']\n" ] } ], "id": "ceca38b2" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 6. Define the Final Feature Set\n", "\n", "Based on the correlation analysis, define the final features used for training:\n", "\n", "**Final feature list:**\n", "- Station features:`station_id`, `capacity`, `lat`, `lon`\n", "- TimeFeature：`hour`, `day`, `day_of_week`, `is_weekend`\n", "- Weather features:`avg_temperature`, `avg_humidity`, `avg_pressure`\n", "\n", "**Notes:**\n", "- Remove original weather columns (`max/min`) and keep average versions to reduce redundancy\n", "- Keep all station and time features\n", "- Target variable: `num_bikes_available`" ], "id": "ad82f535" }, { "cell_type": "code", "metadata": {}, "source": [ "# Define features and target variable\n", "features = [\n", " 'station_id', 'capacity', 'lat', 'lon',\n", " 'hour', 'day', 'day_of_week', 'is_weekend',\n", " 'avg_temperature', 'avg_humidity', 'avg_pressure'\n", "]\n", "\n", "target = 'num_bikes_available'\n", "\n", "X = df_model[features]\n", "y = df_model[target]\n", "\n", "# Split into train and test sets (70% train, 30% test)\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y, test_size=0.3, random_state=42\n", ")\n", "\n", "print(f\"✅ Dataset split completed\")\n", "print(f\"📊 Training set size: {X_train.shape[0]} rows\")\n", "print(f\"📊 Test set size: {X_test.shape[0]} rows\")\n", "print(f\"📋 Number of features: {X_train.shape[1]} \")\n", "print(f\"\\nFeature list: {features}\")" ], "execution_count": 49, "outputs": [ { "output_type": "stream", "text": [ "✅ Dataset split completed\n", "📊 Training set size: 209262 rows\n", "📊 Test set size: 89684 rows\n", "📋 Number of features: 11 \n", "\n", "Feature list: ['station_id', 'capacity', 'lat', 'lon', 'hour', 'day', 'day_of_week', 'is_weekend', 'avg_temperature', 'avg_humidity', 'avg_pressure']\n" ] } ], "id": "68a7ae8f" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 7. Model Training and Comparison\n", "\n", "We will train the following four regression models and compare their performance:\n", "\n", "| Model | Description |\n", "|------|------|\n", "| Linear Regression | Linear regression as a baseline model |\n", "| Decision Tree | Decision tree, can capture nonlinear relationships |\n", "| Random Forest | Random forest, ensemble learning, usually performs better |\n", "| Gradient Boosting | Gradient boosting, often the strongest traditional ML model |\n", "\n", "**Evaluation metrics:**\n", "- `MAE` (Mean Absolute Error): lower is better\n", "- `RMSE` (Root Mean Squared Error): lower is better \n", "- `R²` (Coefficient of Determination): closer to 1 is better" ], "id": "1cab8aa8" }, { "cell_type": "code", "metadata": {}, "source": [ "from sklearn.metrics import mean_squared_error\n", "\n", "# Define models\n", "models = {\n", " 'Linear Regression': LinearRegression(),\n", " 'Decision Tree': DecisionTreeRegressor(random_state=42),\n", " 'Random Forest': RandomForestRegressor(n_estimators=100, random_state=42, n_jobs=-1, max_depth=15, min_samples_leaf=10),\n", " 'Gradient Boosting': GradientBoostingRegressor(n_estimators=100, random_state=42)\n", "}\n", "\n", "# Train and evaluate each model\n", "results = []\n", "\n", "for name, model in models.items():\n", " print(f\"⏳ Training {name}...\")\n", " \n", " # Train\n", " model.fit(X_train, y_train)\n", " \n", " # Predict\n", " y_pred = model.predict(X_test)\n", " \n", " # Evaluate\n", " mae = mean_absolute_error(y_test, y_pred)\n", " rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n", " r2 = r2_score(y_test, y_pred)\n", " \n", " results.append({\n", " 'Model': name,\n", " 'MAE': round(mae, 4),\n", " 'RMSE': round(rmse, 4),\n", " 'R²': round(r2, 4)\n", " })\n", " \n", " print(f\" ✅ MAE: {mae:.4f} | RMSE: {rmse:.4f} | R²: {r2:.4f}\")\n", "\n", "# Summarize results\n", "results_df = pd.DataFrame(results).set_index('Model')\n", "print(f\"\\n📊 Model comparison summary:\")\n", "print(results_df)" ], "execution_count": 50, "outputs": [ { "output_type": "stream", "text": [ "⏳ Training Linear Regression...\n", " ✅ MAE: 7.8289 | RMSE: 9.3659 | R²: 0.0746\n", "⏳ Training Decision Tree...\n", " ✅ MAE: 0.9700 | RMSE: 2.3755 | R²: 0.9405\n", "⏳ Training Random Forest...\n", " ✅ MAE: 2.3249 | RMSE: 3.3927 | R²: 0.8786\n", "⏳ Training Gradient Boosting...\n", " ✅ MAE: 6.1196 | RMSE: 7.5375 | R²: 0.4007\n", "\n", "📊 Model comparison summary:\n", " MAE RMSE R²\n", "Model \n", "Linear Regression 7.829 9.366 0.075\n", "Decision Tree 0.970 2.376 0.941\n", "Random Forest 2.325 3.393 0.879\n", "Gradient Boosting 6.120 7.537 0.401\n" ] } ], "id": "838e85bc" }, { "cell_type": "code", "metadata": {}, "source": [ "fig, axes = plt.subplots(1, 3, figsize=(14, 5))\n", "\n", "metrics = ['MAE', 'RMSE', 'R²']\n", "colors = ['steelblue', 'tomato', 'seagreen', 'orange']\n", "\n", "for i, metric in enumerate(metrics):\n", " axes[i].bar(results_df.index, results_df[metric], color=colors)\n", " axes[i].set_title(f'{metric} Comparison')\n", " axes[i].set_ylabel(metric)\n", " axes[i].set_xticklabels(results_df.index, rotation=20, ha='right')\n", " \n", " # Display values above bars\n", " for j, v in enumerate(results_df[metric]):\n", " axes[i].text(j, v + 0.01, str(v), ha='center', fontsize=9)\n", "\n", "plt.suptitle('Performance Comparison Across Models', fontsize=14, fontweight='bold')\n", "plt.tight_layout()\n", "plt.show()" ], "execution_count": 51, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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", "text/plain": [ "
" ] } } ], "id": "9828e9d7" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 8. Feature Importance Analysis\n", "\n", "Use the best model (Decision Tree) to inspect feature importance,\n", "and understand which features contribute most to prediction." ], "id": "6ed02544" }, { "cell_type": "code", "metadata": {}, "source": [ "# Get Decision Tree feature importance\n", "best_model = models['Decision Tree']\n", "feature_importance = pd.Series(\n", " best_model.feature_importances_,\n", " index=features\n", ").sort_values(ascending=False)\n", "\n", "print(\"📊 Feature importance ranking:\")\n", "print(feature_importance)\n", "\n", "# Visualization\n", "plt.figure(figsize=(10, 6))\n", "colors = ['steelblue' if i < 3 else 'lightsteelblue' for i in range(len(feature_importance))]\n", "feature_importance.plot(kind='bar', color=colors)\n", "plt.title('Decision Tree Feature Importance')\n", "plt.xlabel('Feature')\n", "plt.ylabel('Importance Score')\n", "plt.xticks(rotation=45, ha='right')\n", "plt.tight_layout()\n", "plt.show()" ], "execution_count": 52, "outputs": [ { "output_type": "stream", "text": [ "📊 Feature importance ranking:\n", "lat 0.224\n", "lon 0.146\n", "day 0.137\n", "hour 0.105\n", "station_id 0.096\n", "avg_pressure 0.077\n", "capacity 0.076\n", "day_of_week 0.061\n", "avg_temperature 0.036\n", "avg_humidity 0.030\n", "is_weekend 0.014\n", "dtype: float64\n" ] }, { "output_type": "display_data", "data": { "image/png": "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", "text/plain": [ "
" ] } } ], "id": "2587e00a" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 9. Save the Best Model\n", "\n", "Save the Decision Tree model as a `.pkl` file\n", "for later loading in the Flask app.\n", "\n", "Also save the feature list to ensure the same feature order is used during prediction." ], "id": "8b344acb" }, { "cell_type": "code", "metadata": {}, "source": [ "# Save Decision Tree model\n", "model_filename = 'bike_availability_model.pkl'\n", "\n", "with open(model_filename, 'wb') as f:\n", " pickle.dump(best_model, f)\n", "\n", "# Save the feature list as well\n", "features_filename = 'model_features.pkl'\n", "with open(features_filename, 'wb') as f:\n", " pickle.dump(features, f)\n", "\n", "print(f\"✅ Model saved to: {model_filename}\")\n", "print(f\"✅ Feature list saved to: {features_filename}\")\n", "\n", "# Validation: reload the model and test\n", "with open(model_filename, 'rb') as f:\n", " loaded_model = pickle.load(f)\n", "\n", "# Validate with the first row of the test set\n", "test_sample = X_test.iloc[[0]]\n", "prediction = loaded_model.predict(test_sample)\n", "actual = y_test.iloc[0]\n", "\n", "print(f\"\\n🔍 Model validation:\")\n", "print(f\" Input features: {test_sample.values[0]}\")\n", "print(f\" Predicted value: {prediction[0]:.2f} bikes\")\n", "print(f\" Actual value: {actual} bikes\")\n", "print(f\" Error: {abs(prediction[0] - actual):.2f} bikes\")" ], "execution_count": 53, "outputs": [ { "output_type": "stream", "text": [ "✅ Model saved to: bike_availability_model.pkl\n", "✅ Feature list saved to: model_features.pkl\n", "\n", "🔍 Model validation:\n", " Input features: [ 97. 40. 53.342113 -6.310015 17. 31.\n", " 1. 0. 9.67 87.4 998.525 ]\n", " Predicted value: 24.99 bikes\n", " Actual value: 31 bikes\n", " Error: 6.01 bikes\n" ] } ], "id": "3b3ce717" }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 10. Summary\n", "\n", "### Model Training Results\n", "\n", "| Model | MAE | RMSE | R² |\n", "|------|-----|------|----|\n", "| Linear Regression | 7.829 | 9.366 | 0.075 |\n", "| **Decision Tree** | **0.970** | **2.376** | **0.941** |\n", "| Random Forest | 2.325 | 3.393 | 0.879 |\n", "| Gradient Boosting | 6.120 | 7.537 | 0.401 |\n", "\n", "### Best Model: Decision Tree\n", "- MAE = 0.970 and RMSE = 2.376, giving the lowest prediction errors among all tested models\n", "- R² = 0.941, indicating strong explanatory power on the current test set\n", "\n", "### Most Important Features\n", "1. `lat` / `lon` — Geographic location\n", "2. `day` / `hour` — Time\n", "3. `avg_pressure` — Pressure\n", "4. `station_id` / `capacity` — Station information\n", "\n", "### Next Steps\n", "- Deploy `bike_availability_model.pkl` to the Flask application\n", "- Use OpenWeather API real-time weather data for prediction" ], "id": "865bd00b" } ], "metadata": { "kernelspec": { "display_name": "se", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.12.12" } }, "nbformat": 4, "nbformat_minor": 5 }