S15 T01 P1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d8a3a070",
   "metadata": {},
   "source": [
    "# Sprint 15, Task 1\n",
    "\n",
    "## Level 1\n",
    "\n",
    "### Exercise 1\n",
    "\n",
    "Classify the different flights using the K-means algorithm:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cc479567",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:48.321134Z",
     "start_time": "2021-07-24T17:36:43.888913Z"
    }
   },
   "outputs": [],
   "source": [
    "# Import useful libraries:\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import sklearn\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_score\n",
    "\n",
    "from kneed import KneeLocator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "43678973",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.163847Z",
     "start_time": "2021-07-24T17:36:48.332442Z"
    }
   },
   "outputs": [],
   "source": [
    "# Open Dataset:\n",
    "\n",
    "dades = pd.read_csv(\"../Sprint 11/DelayedFlights.csv\")\n",
    "\n",
    "# Feature filtering: first we eliminate columns that do not tell us much information about flight delays\n",
    "\n",
    "dades.drop(columns=[\"Unnamed: 0\", \"Year\", \"FlightNum\", \"TailNum\", \"Cancelled\", \"CancellationCode\"], inplace=True)\n",
    "\n",
    "# Since we are going to use PCA, we only want continous numerical values, which will gave maximum explanation potential\n",
    "\n",
    "# Then we filter out variables that are not continuous numerical:\n",
    "\n",
    "dades.drop(columns=[\"Month\", \"DayofMonth\", \"DayOfWeek\", \"UniqueCarrier\", \"Origin\", \"Dest\", \"Diverted\"], inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "33013c83",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.173271Z",
     "start_time": "2021-07-24T17:36:56.169161Z"
    }
   },
   "outputs": [],
   "source": [
    "##################################################################################################################\n",
    "###################################### FILLING IN NULL VALUES ####################################################"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6658e2a4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.244699Z",
     "start_time": "2021-07-24T17:36:56.180101Z"
    }
   },
   "outputs": [],
   "source": [
    "# Arrival Delay: NaN cases filled with Departure Delay value\n",
    "\n",
    "dades[\"ArrDelay\"].fillna(dades[\"DepDelay\"], inplace = True)\n",
    "\n",
    "# Arrival Time: NaN cases filled with Scheduled Arrival Time + Arrival Delay\n",
    "\n",
    "foo = dades[\"CRSArrTime\"] + dades[\"ArrDelay\"]\n",
    "\n",
    "dades[\"ArrTime\"].fillna(foo, inplace = True)           \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "35396a49",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.495887Z",
     "start_time": "2021-07-24T17:36:56.247332Z"
    }
   },
   "outputs": [],
   "source": [
    "# Scheduled Elapsed Time: NaN values filled with Scheduled Arrival Time - Scheduled Departure Time\n",
    "\n",
    "    # Function that will calculate new value:\n",
    "    \n",
    "def calc_time(first, second):\n",
    "    result = first - second\n",
    "    for i in result:\n",
    "        if i < 0:\n",
    "            i =+ 2400\n",
    "    return result\n",
    "\n",
    "dades[\"CRSElapsedTime\"].fillna(calc_time(dades[\"CRSArrTime\"], dades[\"CRSDepTime\"]), inplace=True)\n",
    "\n",
    "# Actual Elapsed Time: NaN values filled with Scheduled Elapsed Time + Arrival Delay\n",
    "\n",
    "calculate = dades[\"CRSElapsedTime\"] + dades[\"ArrDelay\"]\n",
    "\n",
    "dades[\"ActualElapsedTime\"].fillna(calculate, inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5b1d6495",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.542542Z",
     "start_time": "2021-07-24T17:36:56.498187Z"
    }
   },
   "outputs": [],
   "source": [
    "# Taxi In and Out: NaN values filled in with column mean\n",
    "\n",
    "dades[\"TaxiIn\"].fillna(dades[\"TaxiIn\"].mean(), inplace=True)\n",
    "\n",
    "dades[\"TaxiOut\"].fillna(dades[\"TaxiOut\"].mean(), inplace=True)\n",
    "\n",
    "# Air Time: NaN values filled with Actual Elapsed Time - Taxi Time\n",
    "\n",
    "fum = dades[\"ActualElapsedTime\"] - dades[\"TaxiIn\"] - dades[\"TaxiOut\"]\n",
    "\n",
    "dades[\"AirTime\"].fillna(fum, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8c9dba9f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.635685Z",
     "start_time": "2021-07-24T17:36:56.546799Z"
    }
   },
   "outputs": [],
   "source": [
    "# Specific reasons for Delay: filled in with 0\n",
    "\n",
    "dades[\"CarrierDelay\"].fillna(0, inplace=True)\n",
    "\n",
    "dades[\"WeatherDelay\"].fillna(0, inplace=True)\n",
    "\n",
    "dades[\"NASDelay\"].fillna(0, inplace=True)\n",
    "\n",
    "dades[\"SecurityDelay\"].fillna(0, inplace=True)\n",
    "\n",
    "dades[\"LateAircraftDelay\"].fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2774643a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:56.713008Z",
     "start_time": "2021-07-24T17:36:56.640956Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DepTime              0\n",
       "CRSDepTime           0\n",
       "ArrTime              0\n",
       "CRSArrTime           0\n",
       "ActualElapsedTime    0\n",
       "CRSElapsedTime       0\n",
       "AirTime              0\n",
       "ArrDelay             0\n",
       "DepDelay             0\n",
       "Distance             0\n",
       "TaxiIn               0\n",
       "TaxiOut              0\n",
       "CarrierDelay         0\n",
       "WeatherDelay         0\n",
       "NASDelay             0\n",
       "SecurityDelay        0\n",
       "LateAircraftDelay    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dades.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "965ba6fb",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:36:57.532045Z",
     "start_time": "2021-07-24T17:36:56.717585Z"
    }
   },
   "outputs": [],
   "source": [
    "# Scale from 0 to 1:\n",
    "\n",
    "scaler = MinMaxScaler()\n",
    "\n",
    "scaled_df = scaler.fit_transform(dades)\n",
    "\n",
    "scaled_df = pd.DataFrame(scaled_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5e97b19b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:37:03.936858Z",
     "start_time": "2021-07-24T17:36:57.534503Z"
    }
   },
   "outputs": [],
   "source": [
    "# PCA transformation to 2 elements:\n",
    "\n",
    "pca = PCA(n_components=2)\n",
    "\n",
    "principalComponents = pca.fit_transform(scaled_df)\n",
    "\n",
    "new_df = pd.DataFrame(data = principalComponents)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5ca06814",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:37:06.032015Z",
     "start_time": "2021-07-24T17:37:03.939384Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7faa4f06cfa0>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# See scatterplot after PCA transformation: \n",
    "\n",
    "plt.scatter(new_df[0], new_df[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4f5f8767",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:37:28.496599Z",
     "start_time": "2021-07-24T17:37:06.046136Z"
    }
   },
   "outputs": [],
   "source": [
    "# K-means algorithm: set hyperparameters, fit and predict; intuitively choose 6 clusters given scatterplot shape\n",
    "\n",
    "kmeans = KMeans(init=\"random\", n_clusters=6, random_state=59)\n",
    "\n",
    "labels = kmeans.fit_predict(new_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "facc18a0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:37:31.262345Z",
     "start_time": "2021-07-24T17:37:28.499383Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Getting array with labels:\n",
    " \n",
    "u_labels = np.unique(labels)\n",
    " \n",
    "# Then plotting the results:\n",
    " \n",
    "for i in u_labels:\n",
    "    filtered_label = new_df[labels == i]\n",
    "    plt.scatter(filtered_label[0] , filtered_label[1])\n",
    "\n",
    "# It seems to have found groups on the extremities of the figure (top, right and left), but the central groupings\n",
    "# look unintuitive. In fact, when changing the number of centroids those clusters remain (conceptually) unchanged."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c21ab4ba",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:37:31.282860Z",
     "start_time": "2021-07-24T17:37:31.265888Z"
    }
   },
   "outputs": [],
   "source": [
    "# Silhouette has not been applied because of its high computational burden; we will try elbow method instead"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "db103eac",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:38.054518Z",
     "start_time": "2021-07-24T17:37:31.287912Z"
    }
   },
   "outputs": [],
   "source": [
    "# Elbow method calculations:\n",
    "\n",
    "sse = []\n",
    "\n",
    "for k in range(1, 11):\n",
    "    kmeans = KMeans(n_clusters=k, max_iter= 300, random_state= 88)\n",
    "    kmeans.fit(new_df)\n",
    "    sse.append(kmeans.inertia_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "73a00af0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:38.272824Z",
     "start_time": "2021-07-24T17:41:38.057150Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot results:\n",
    "\n",
    "plt.style.use(\"fivethirtyeight\")\n",
    "plt.plot(range(1, 11), sse)\n",
    "plt.xticks(range(1, 11))\n",
    "plt.xlabel(\"Number of Clusters\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.show()\n",
    "\n",
    "# Elbow method shows 2 or 4 might be the best result. Let us use Knee Locator instead for a clearer insight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "579dec37",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:38.291792Z",
     "start_time": "2021-07-24T17:41:38.276766Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kl = KneeLocator(range(1, 11), sse, curve=\"convex\", direction=\"decreasing\")\n",
    "\n",
    "kl.elbow\n",
    "\n",
    "# 4 is the best number of clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "55ecd965",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:52.207385Z",
     "start_time": "2021-07-24T17:41:38.296504Z"
    }
   },
   "outputs": [],
   "source": [
    "# Now running with 4 clusters:\n",
    "\n",
    "kmeans = KMeans(init=\"random\", n_clusters=4, random_state=14)\n",
    "\n",
    "labels = kmeans.fit_predict(new_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "cdb976ca",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:54.574062Z",
     "start_time": "2021-07-24T17:41:52.210323Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Now we plot:\n",
    "\n",
    "u_labels = np.unique(labels)\n",
    " \n",
    "for i in u_labels:\n",
    "    filtered_label = new_df[labels == i]\n",
    "    plt.scatter(filtered_label[0] , filtered_label[1])\n",
    "    \n",
    "# It looks fundamentally better in theoretical terms than 6 clusters model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d046a684",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:54.581919Z",
     "start_time": "2021-07-24T17:41:54.576196Z"
    }
   },
   "outputs": [],
   "source": [
    "# For the continuation of the exercise, see S15 T01 P2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "8c345703",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-07-24T17:41:54.594015Z",
     "start_time": "2021-07-24T17:41:54.585298Z"
    }
   },
   "outputs": [],
   "source": [
    "###################################################################################################################"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}