Chicago Taxi OLAP.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import division\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from datetime import datetime\n",
    "from dateutil.relativedelta import relativedelta as rd\n",
    "from sqlalchemy import create_engine \n",
    "from analytics_tools.analytics_tools3 import freq_discrete"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "conn_oltp= create_engine('mysql://root:123@localhost:3306/taxi_oltp').connect()\n",
    "conn_olap= create_engine('mysql://root:123@localhost:3306/taxi_olap').connect()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TD_company "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"\"\"\n",
    "SELECT \n",
    "    C.company, C.comp_name, A.id_trip\n",
    "FROM\n",
    "    tbl_trip A\n",
    "        INNER JOIN\n",
    "    tbl_taxi B ON A.taxi_id = B.taxi_id\n",
    "        INNER JOIN\n",
    "    tbl_company C ON B.company = C.company\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "df = pd.read_sql(sql=query,con=conn_oltp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux = df[['company','comp_name','id_trip']].groupby(['company','comp_name']).count().sort_values('id_trip',ascending=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux.reset_index(inplace=True)\n",
    "aux['comp_name_top10'] = np.where(aux.index<10,aux.comp_name,'otro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux[['company','comp_name_top10']].to_sql(chunksize=1000,\n",
    "         con=conn_olap,\n",
    "         if_exists='replace',\n",
    "         name='td_company',\n",
    "         index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TD_mop "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"\"\"\n",
    "SELECT \n",
    "    payment_type,\n",
    "    1 as n\n",
    "FROM\n",
    "    tbl_trip\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "df = pd.read_sql(sql=query,con=conn_oltp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux = df.groupby('payment_type').count().sort_values('n',ascending=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux.reset_index(inplace=True)\n",
    "aux['mop'] = np.where(aux.index<2,aux.payment_type,'otro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux.rename(columns={'index':'id_mop'},inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux[['id_mop','payment_type','mop']].to_sql(chunksize=1000,\n",
    "         con=conn_olap,\n",
    "         if_exists='replace',\n",
    "         name='td_mop',\n",
    "         index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TD_tiempo "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"\"\"\n",
    "SELECT \n",
    "    trip_start_timestamp,\n",
    "    trip_end_timestamp\n",
    "FROM\n",
    "    tbl_trip2\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "df = pd.read_sql(sql=query,con=conn_oltp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.to_datetime(df.trip_start_timestamp).min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fhi = datetime(2016,1,1,0,0)\n",
    "fhf = datetime(2020,12,31,23,30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "lst_fh = []\n",
    "while fhi<fhf:\n",
    "    for _ in range(24):\n",
    "        fhi+=rd(hours=1)\n",
    "        lst_fh.append(fhi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(lst_fh,columns=['id_fh'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "datetime.now().weekday()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['year'] = df.id_fh.map(lambda x:x.year)\n",
    "df['month'] = df.id_fh.map(lambda x:x.month)\n",
    "df['day'] = df.id_fh.map(lambda x:x.day)\n",
    "df['hour'] = df.id_fh.map(lambda x:x.hour)\n",
    "df['weekday'] = df.id_fh.map(lambda x:x.weekday())\n",
    "df['id_fh'] = df.id_fh.map(lambda x:x.strftime('%Y%m%d%H'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_sql(chunksize=1000,\n",
    "         con=conn_olap,\n",
    "         if_exists='replace',\n",
    "         name='td_time',\n",
    "         index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TH_trip "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"\"\"\n",
    "    SELECT \n",
    "        A.trip_start_timestamp,\n",
    "        A.fare AS h_fare,\n",
    "        A.tips AS h_tips,\n",
    "        A.tolls AS h_tolls,\n",
    "        A.extras AS h_extras,\n",
    "        A.payment_type,\n",
    "        C.company\n",
    "    FROM\n",
    "        tbl_trip2 A\n",
    "            INNER JOIN\n",
    "        tbl_taxi B ON A.taxi_id = B.taxi_id\n",
    "            INNER JOIN\n",
    "        tbl_company C ON B.company = C.company\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_sql(sql=query,con=conn_oltp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['id_fh'] = pd.to_datetime(df.trip_start_timestamp).map(lambda x:x.strftime('%Y%m%d%H'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['h_total'] = df[[x for x in df.columns if x[:2]=='h_']].sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['h_n'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux = df[['id_fh','payment_type','company']+[x for x in df.columns if x[:2]=='h_']].groupby(['id_fh','payment_type','company']).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "aux.to_sql(chunksize=1000,\n",
    "         con=conn_olap,\n",
    "         if_exists='replace',\n",
    "         name='th_trip',\n",
    "         index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Extracción del Cubo "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "cubo = pd.read_sql(sql=\"select * from v_cubote\",con=conn_olap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20528, 13)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cubo.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "cubo.to_excel('chicago_taxi_cubo_olap.xlsx',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}