airbnb_lisbon_svm.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np #Deal with numbers and arrays\n",
    "import pandas as pd #Create dataframes and sort/clean data faster\n",
    "\n",
    "import matplotlib.pyplot as plt #Visualization module\n",
    "import pathlib as pl\n",
    "import seaborn as sns #Makes matplotlib prettier\n",
    "\n",
    "#This blocks imports everything needed from sklearn(models, procession packages, and metrics)\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score\n",
    "from sklearn.svm import SVR, LinearSVR\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import uniform\n",
    "\n",
    "import plotly.express as px #Module for dynamic data visualization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "latitude                          0\n",
       "longitude                         0\n",
       "room_type                         0\n",
       "accommodates                      0\n",
       "bedrooms                          0\n",
       "beds                              0\n",
       "price                             0\n",
       "minimum_nights                    0\n",
       "number_of_reviews                 0\n",
       "calculated_host_listings_count    0\n",
       "bathrooms                         0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = pl.Path(f\"./data/21\")\n",
    "dfraw21 = pd.DataFrame()\n",
    "for file in path.iterdir():\n",
    "    df= pd.read_csv(path/file.name,index_col=False,usecols=['latitude', 'longitude',\n",
    "       'accommodates', 'bathrooms_text', 'bedrooms', 'beds','price','number_of_reviews',\n",
    "       'calculated_host_listings_count','minimum_nights','room_type'])\n",
    "    df=df.dropna(subset=['price'])\n",
    "    df[\"price\"] = df[\"price\"].apply(lambda x: x.replace(\"$\",\"\"))\n",
    "    df[\"price\"] = df[\"price\"].apply(lambda x: x.replace(\",\",\"\"))\n",
    "    df[\"price\"] = df[\"price\"].astype (np.float32)\n",
    "    df[\"price\"] = df[\"price\"].astype (np.int32)\n",
    "    df['bathrooms'] = df['bathrooms_text'].str.extract(\"(\\d*\\.?\\d+)\", expand=False)\n",
    "    df['bathrooms'] = np.where(df['bathrooms_text'].str.contains(\"half\", case=False, na=False), 0.5, df['bathrooms'])\n",
    "    dfraw21 = pd.concat([dfraw21,df])\n",
    "dfraw21 = dfraw21.dropna()\n",
    "dfraw21 = dfraw21.drop(\"bathrooms_text\", axis = 1)\n",
    "dfraw21.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Treatment\n",
    "See types."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "latitude                          float64\n",
       "longitude                         float64\n",
       "room_type                          object\n",
       "accommodates                        int64\n",
       "bedrooms                          float64\n",
       "beds                              float64\n",
       "price                               int32\n",
       "minimum_nights                      int64\n",
       "number_of_reviews                   int64\n",
       "calculated_host_listings_count      int64\n",
       "bathrooms                          object\n",
       "dtype: object"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfraw21.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Convert 64 int to 32, and object to number"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfraw21[\"accommodates\"] = dfraw21[\"accommodates\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"bedrooms\"] = dfraw21[\"bedrooms\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"beds\"] = dfraw21[\"beds\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"minimum_nights\"] = dfraw21[\"minimum_nights\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"number_of_reviews\"] = dfraw21[\"number_of_reviews\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"calculated_host_listings_count\"] = dfraw21[\"calculated_host_listings_count\"].astype(np.int32, copy = False)\n",
    "dfraw21[\"latitude\"] = dfraw21[\"latitude\"].astype(np.float32, copy = False)\n",
    "dfraw21[\"longitude\"] = dfraw21[\"longitude\"].astype(np.float32, copy = False)\n",
    "dfraw21[\"bathrooms\"] = dfraw21[\"bathrooms\"].astype(np.float32, copy = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remove anomalies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6106 rows removed\n",
      "65503 total rows and 11 total columns\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def outlier (col):\n",
    "    q1 = col.quantile (0.25)\n",
    "    q3 = col.quantile (0.75)\n",
    "    iqr = q3 - q1\n",
    "    lenght = 1.5\n",
    "    return q1 - lenght * iqr, q3 + lenght * iqr\n",
    "\n",
    "def hist (col):\n",
    "    plt.figure (figsize = (15,5))\n",
    "    sns.histplot (col, kde = True)\n",
    "\n",
    "def remove (ds, col):\n",
    "    before = ds.shape [0]\n",
    "    low, up = outlier (ds [col])\n",
    "    ds = ds.loc [(ds [col] >= low) & (ds [col] <= up), :]\n",
    "    return ds, before - ds.shape [0]\n",
    "dfraw21, removed = remove (dfraw21, \"price\")\n",
    "\n",
    "aux = dfraw21 [\"room_type\"].value_counts ()\n",
    "for kind in aux.index:\n",
    "    if aux [kind] < 15000:\n",
    "        dfraw21.loc [dfraw21 [\"room_type\"] == kind, \"room_type\"] = \"Rooms\"\n",
    "        \n",
    "print (f\"{removed} rows removed\")\n",
    "hist (dfraw21 [\"price\"])\n",
    "print (f\"{dfraw21.shape [0]} total rows and {dfraw21.shape [1]} total columns\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfraw21[\"room_type\"] = np.where(dfraw21[\"room_type\"] == \"Rooms\", 0, 1)\n",
    "dfraw21[\"room_type\"] = dfraw21[\"room_type\"].astype(np.int32, copy = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some functions to evaluate methods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'C': 71.21818954206266, 'epsilon': 0.16819429435540822}\n",
      "\n",
      "-----\n",
      "Model: SVR\n",
      "RMSE: 31.34\n",
      "MAE: 23.11\n",
      "R2: 28.94%\n",
      "-----\n",
      "\n",
      "{'n_neighbors': 3}\n",
      "\n",
      "-----\n",
      "Model: KNeighborsRegressor\n",
      "RMSE: 26.83\n",
      "MAE: 18.85\n",
      "R2: 47.92%\n",
      "-----\n",
      "\n"
     ]
    }
   ],
   "source": [
    "def evaluate(model, ytest, prediction):\n",
    "    rmse = np.sqrt (mean_squared_error (ytest, prediction))\n",
    "    mae = mean_absolute_error (ytest, prediction)\n",
    "    r2 = r2_score (ytest, prediction)\n",
    "    return f\"\\n-----\\nModel: {model}\\nRMSE: {rmse:.2f}\\nMAE: {mae:.2f}\\nR2: {r2:.2%}\\n-----\\n\"\n",
    "\n",
    "y = dfraw21['price']\n",
    "columns_dropped = ['price']\n",
    "\n",
    "param_grid_knn = {'n_neighbors': range(1, 21)}\n",
    "\n",
    "param_distributions = {\n",
    "    'C': uniform(0.1, 100),\n",
    "    'epsilon': uniform(0.01, 1)\n",
    "}\n",
    "\n",
    "models = {'SVR': (SVR(), param_distributions),\n",
    "          'KNeighborsRegressor': (KNeighborsRegressor(), param_grid_knn)\n",
    "}\n",
    "\n",
    "X = dfraw21.drop(columns_dropped, axis=1)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=1, train_size=0.8)\n",
    "\n",
    "for name, model in models.items():\n",
    "    random_search = RandomizedSearchCV(model[0], param_distributions=model[1], cv=5)\n",
    "    random_search.fit(X_train, y_train)\n",
    "    predictions = random_search.predict(X_test)\n",
    "    print(random_search.best_params_)\n",
    "    print(evaluate(name, y_test, predictions))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}