{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizing data for analysis with Amazon Athena and AWS Glue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will continue our open data analytics workflow starting with the AWS Console then moving to using the notebook. Using [AWS Glue](https://aws.amazon.com/glue/) we can automate creating a metadata catalog based on flat files stored on Amazon S3. Glue is a fully managed extract, transform, and load (ETL) service that makes it easy for customers to prepare and load their data for analytics. You can create and run an ETL job with a few clicks in the AWS Management Console. You simply point AWS Glue to your data stored on AWS, and AWS Glue discovers your data and stores the associated metadata (e.g. table definition and schema) in the AWS Glue Data Catalog. Once cataloged, your data is immediately searchable, queryable, and available for ETL.\n",
    "\n",
    "### Glue Data Catalog\n",
    "\n",
    "We have sourced the open dataset from the [Registry of Open Data on AWS](https://registry.opendata.aws/). We also stored the data on S3. Now we are ready to extract, transform, and load the data for analytics. We will use AWS Glue service to do this. First step is to create a logical database entry in the data catalog. Note that we are not creating a physical database which requires resources. This is just a metadata placeholder for the flat file we copied into S3.\n",
    "\n",
    "> When creating the data catalog name try choosing a name without hyphens and few characters long. This will make SQL queries more readable and also avoid certain errors when running these queries.\n",
    "\n",
    "![Glue Data Catalog](https://s3.amazonaws.com/cloudstory/notebooks-media/glue-data-catalog.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also setup the notebook for accessing AWS Glue service using the ``Boto3`` Python SDK. The ``pandas`` and ``IPython`` dependencies are imported for output formatting purposes only. We also import ``numpy`` a popular statistical analysis library. Charts and visualizations will be supported by ``seaborn`` and ``matplotlib`` libraries. To access the Glue service API we create a Glue client. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import boto3\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from IPython.display import display, Markdown\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "glue = boto3.client('glue')\n",
    "s3 = boto3.client('s3')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### List Glue Databases\n",
    "We will recreate the AWS Console GUI experience using SDK calls by creating the ``list_glue_databases`` function. We simply get the data catalogs in one statement and iterate over the results in the next one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def list_glue_databases():\n",
    "    glue_database = glue.get_databases()\n",
    "\n",
    "    for db in glue_database['DatabaseList']:\n",
    "        print(db['Name'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "default\n",
      "odoc\n",
      "sampledb\n",
      "taxicatalog\n"
     ]
    }
   ],
   "source": [
    "list_glue_databases()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Glue Crawler\n",
    "Next, we create a logical table using Glue crawler. This is again just table metadata definition while the actual data is still stored only in the flat file on S3. For this notebook we will define and run the default Glue Crawler to extract and load the metadata schema from our flat file. This requires selection of a data store which is S3 in this case, defining an IAM role for access from Glue to S3, selecting a schedule for the crawler to run repeatedly if required, and output destination of the crawler results. \n",
    "\n",
    "> Please ensure that the flat file is stored on S3 within its own folder and you point at the folder when picking the data source during crawler definition. If you point directly to a flat file when running the crawler, it may return zero results when querying using Amazon Athena.\n",
    "\n",
    "Glue will pick up folder name for the logical table name. Keeping our data source files in a folder has the added advantage of incremntally updating the folder with updates to the data with more files or updating the original file. Glue will pick up these changes based on crawler run schedule.\n",
    "\n",
    "![Glue Crawler](https://s3.amazonaws.com/cloudstory/notebooks-media/glue-crawler.png)\n",
    "\n",
    "### Glue Table Metadata\n",
    "This results in extraction of table metadata stored within our data catalog. The schema with data types is extracted and stored in Glue Data Catalog. Note that the default Glue Crawler understands well-formed CSV files with first row as comma-separated list of column names, and next set of rows representing ordered data records. The Glue Crawler automatically guesses data types based on the contents of the flat file.\n",
    "\n",
    "![Table Metadata](https://s3.amazonaws.com/cloudstory/notebooks-media/table-metadata.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transform Data Using Athena\n",
    "\n",
    "Transforming big data in notebook environment is not viable. Instead we can use Amazon Athena for large data transforms and bring the results back into our notebook.\n",
    "\n",
    "![Athena Transform Data](https://s3.amazonaws.com/cloudstory/notebooks-media/athena-transform-data.png)\n",
    "\n",
    "We will use following query to create a well formed table transformed from our original table. Note that we specify the output location so that Athena defined WorkGroup location is not used by default. We also specify the format as ``TEXTFILE`` otherwise default ``PARQUET`` format is used which may generate errors when sampling this data.\n",
    "\n",
    "```SQL\n",
    "CREATE TABLE \n",
    "IF NOT EXISTS \"taxicatalog\".\"many_trips_well_formed\" \n",
    "WITH (\n",
    "    external_location = 's3://open-data-analytics-taxi-trips/many-trips-well-formed/',\n",
    "    format = 'TEXTFILE',\n",
    "    field_delimiter = ','\n",
    ")\n",
    "AS SELECT vendorid AS vendor,\n",
    "         passenger_count AS passengers,\n",
    "         trip_distance AS distance,\n",
    "         ratecodeid AS rate,\n",
    "         pulocationid AS pick_location,\n",
    "         dolocationid AS drop_location,\n",
    "         payment_type AS payment_type,\n",
    "         fare_amount AS fare,\n",
    "         extra AS extra_fare,\n",
    "         mta_tax AS tax,\n",
    "         tip_amount AS tip,\n",
    "         tolls_amount AS toll,\n",
    "         improvement_surcharge AS surcharge,\n",
    "         total_amount AS total_fare,\n",
    "         tpep_pickup_datetime AS pick_when,\n",
    "         tpep_dropoff_datetime AS drop_when\n",
    "FROM \"taxicatalog\".\"many_trips\";\n",
    "```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### List Glue Tables\n",
    "In the spirit of AWS Open Data Analytics API we will recreate the AWS Console feature which lists the tables and displays the metadata within one single reusable function. We get the list of table metadata stored within our data catalog by passing the ``database`` parameter. Next we iterate over each table object and display the name, source data file, number of records (estimate), average record size, data size in MB, and the name of the crawler used to extract the table metadata. We also display the list of column names and data types extracted as schema from the flat file stored on S3."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def list_glue_tables(database, verbose=True):\n",
    "    glue_tables = glue.get_tables(DatabaseName=database)\n",
    "    \n",
    "    for table in glue_tables['TableList']:\n",
    "        display(Markdown('**Table: ' + table['Name'] + '**'))\n",
    "        display(Markdown('Location: ' + table['StorageDescriptor']['Location']))\n",
    "        created = table['CreatedBy'].split('/')\n",
    "        display(Markdown('Created by: ' + created[-1]))\n",
    "        if verbose and created[-1] == 'AWS Crawler':\n",
    "            display(Markdown(f'Records: {int(table[\"Parameters\"][\"recordCount\"]):,}'))\n",
    "            display(Markdown(f'Average Record Size: {table[\"Parameters\"][\"averageRecordSize\"]} Bytes'))\n",
    "            display(Markdown(f'Dataset Size: {float(table[\"Parameters\"][\"sizeKey\"])/1024/1024:3.0f} MB'))\n",
    "            display(Markdown(f'Crawler: {table[\"Parameters\"][\"UPDATED_BY_CRAWLER\"]}'))\n",
    "        if verbose:\n",
    "            df_columns = pd.DataFrame.from_dict(table[\"StorageDescriptor\"][\"Columns\"])\n",
    "            display(df_columns[['Name', 'Type']])\n",
    "            display(Markdown('---'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Table: many_trips**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "Location: s3://open-data-analytics-taxi-trips/many-trips/"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "Created by: AWS-Crawler"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "**Table: many_trips_well_formed**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "Location: s3://open-data-analytics-taxi-trips/many-trips-well-formed"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "Created by: manav"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "list_glue_tables('taxicatalog', verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "athena = boto3.client('athena')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Athena Query\n",
    "Our next action is to bring the data created within Athena into the notebook environment using a ``pandas`` DataFrame. This can be done using the ``athena_query`` function which calls the Amazon Athena API to execute a query and store the output within a bucket and folder. This output is then read by a DataFrame which is returned by the function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def athena_query(query, bucket, folder):\n",
    "    output = 's3://' + bucket + '/' + folder + '/'\n",
    "    response = athena.start_query_execution(QueryString=query, \n",
    "                                        ResultConfiguration={'OutputLocation': output})\n",
    "    qid = response['QueryExecutionId']\n",
    "    response = athena.get_query_execution(QueryExecutionId=qid)\n",
    "    state = response['QueryExecution']['Status']['State']\n",
    "    while state == 'RUNNING':\n",
    "        response = athena.get_query_execution(QueryExecutionId=qid)\n",
    "        state = response['QueryExecution']['Status']['State']\n",
    "    key = folder + '/' + qid + '.csv'\n",
    "    data_source = {'Bucket': bucket, 'Key': key}\n",
    "    url = s3.generate_presigned_url(ClientMethod = 'get_object', Params = data_source)\n",
    "    data = pd.read_csv(url)\n",
    "    return data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To explore the data within Athena we will query returning thousand random samples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>vendor</th>\n",
       "      <th>passengers</th>\n",
       "      <th>distance</th>\n",
       "      <th>rate</th>\n",
       "      <th>pick_location</th>\n",
       "      <th>drop_location</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>fare</th>\n",
       "      <th>extra_fare</th>\n",
       "      <th>tax</th>\n",
       "      <th>tip</th>\n",
       "      <th>toll</th>\n",
       "      <th>surcharge</th>\n",
       "      <th>total_fare</th>\n",
       "      <th>pick_when</th>\n",
       "      <th>drop_when</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1.25</td>\n",
       "      <td>1</td>\n",
       "      <td>237</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.3</td>\n",
       "      <td>9.80</td>\n",
       "      <td>2018-06-06 10:43:34</td>\n",
       "      <td>2018-06-06 10:54:58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1.20</td>\n",
       "      <td>1</td>\n",
       "      <td>158</td>\n",
       "      <td>90</td>\n",
       "      <td>2</td>\n",
       "      <td>7.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.3</td>\n",
       "      <td>8.30</td>\n",
       "      <td>2018-06-06 10:06:22</td>\n",
       "      <td>2018-06-06 10:15:21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3.30</td>\n",
       "      <td>1</td>\n",
       "      <td>234</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>3.55</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.3</td>\n",
       "      <td>21.35</td>\n",
       "      <td>2018-06-06 10:17:20</td>\n",
       "      <td>2018-06-06 10:43:07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.90</td>\n",
       "      <td>1</td>\n",
       "      <td>236</td>\n",
       "      <td>140</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1.55</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.3</td>\n",
       "      <td>9.35</td>\n",
       "      <td>2018-06-06 10:48:28</td>\n",
       "      <td>2018-06-06 10:57:08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1</td>\n",
       "      <td>141</td>\n",
       "      <td>162</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.3</td>\n",
       "      <td>9.75</td>\n",
       "      <td>2018-06-06 10:59:28</td>\n",
       "      <td>2018-06-06 11:08:05</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   vendor  passengers  distance  rate  pick_location  drop_location  \\\n",
       "0       2           1      1.25     1            237            236   \n",
       "1       1           1      1.20     1            158             90   \n",
       "2       1           1      3.30     1            234            236   \n",
       "3       1           1      0.90     1            236            140   \n",
       "4       1           1      1.00     1            141            162   \n",
       "\n",
       "   payment_type  fare  extra_fare  tax   tip  toll  surcharge  total_fare  \\\n",
       "0             1   9.0         0.0  0.5  0.00   0.0        0.3        9.80   \n",
       "1             2   7.5         0.0  0.5  0.00   0.0        0.3        8.30   \n",
       "2             1  17.0         0.0  0.5  3.55   0.0        0.3       21.35   \n",
       "3             1   7.0         0.0  0.5  1.55   0.0        0.3        9.35   \n",
       "4             1   7.0         0.0  0.5  1.95   0.0        0.3        9.75   \n",
       "\n",
       "             pick_when            drop_when  \n",
       "0  2018-06-06 10:43:34  2018-06-06 10:54:58  \n",
       "1  2018-06-06 10:06:22  2018-06-06 10:15:21  \n",
       "2  2018-06-06 10:17:20  2018-06-06 10:43:07  \n",
       "3  2018-06-06 10:48:28  2018-06-06 10:57:08  \n",
       "4  2018-06-06 10:59:28  2018-06-06 11:08:05  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bucket = 'open-data-analytics-taxi-trips'\n",
    "folder = 'queries'\n",
    "query = 'SELECT * FROM \"taxicatalog\".\"many_trips_well_formed\" TABLESAMPLE BERNOULLI(100) LIMIT 1000;'\n",
    "\n",
    "df = athena_query(query, bucket, folder)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next we will determine statistical correlation between various features (columns) within the given set of samples (records)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>vendor</th>\n",
       "      <th>passengers</th>\n",
       "      <th>distance</th>\n",
       "      <th>rate</th>\n",
       "      <th>pick_location</th>\n",
       "      <th>drop_location</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>fare</th>\n",
       "      <th>extra_fare</th>\n",
       "      <th>tax</th>\n",
       "      <th>tip</th>\n",
       "      <th>toll</th>\n",
       "      <th>surcharge</th>\n",
       "      <th>total_fare</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>vendor</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.283619</td>\n",
       "      <td>0.015401</td>\n",
       "      <td>0.055897</td>\n",
       "      <td>-0.024097</td>\n",
       "      <td>-0.005115</td>\n",
       "      <td>0.013634</td>\n",
       "      <td>-0.014083</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.009308</td>\n",
       "      <td>-0.024436</td>\n",
       "      <td>0.025840</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.015078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>passengers</th>\n",
       "      <td>0.283619</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.033053</td>\n",
       "      <td>0.051624</td>\n",
       "      <td>-0.021166</td>\n",
       "      <td>0.003783</td>\n",
       "      <td>0.020200</td>\n",
       "      <td>0.035106</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.022736</td>\n",
       "      <td>0.008936</td>\n",
       "      <td>0.003765</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.033289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distance</th>\n",
       "      <td>0.015401</td>\n",
       "      <td>0.033053</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.119010</td>\n",
       "      <td>-0.119491</td>\n",
       "      <td>-0.148011</td>\n",
       "      <td>-0.068732</td>\n",
       "      <td>0.917127</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.080828</td>\n",
       "      <td>0.389773</td>\n",
       "      <td>0.401863</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.903529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rate</th>\n",
       "      <td>0.055897</td>\n",
       "      <td>0.051624</td>\n",
       "      <td>0.119010</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.042557</td>\n",
       "      <td>-0.053956</td>\n",
       "      <td>-0.007774</td>\n",
       "      <td>0.185992</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.501256</td>\n",
       "      <td>0.083778</td>\n",
       "      <td>0.246460</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.184445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pick_location</th>\n",
       "      <td>-0.024097</td>\n",
       "      <td>-0.021166</td>\n",
       "      <td>-0.119491</td>\n",
       "      <td>-0.042557</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.150656</td>\n",
       "      <td>-0.009998</td>\n",
       "      <td>-0.129692</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.010869</td>\n",
       "      <td>-0.028087</td>\n",
       "      <td>-0.153488</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.127936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>drop_location</th>\n",
       "      <td>-0.005115</td>\n",
       "      <td>0.003783</td>\n",
       "      <td>-0.148011</td>\n",
       "      <td>-0.053956</td>\n",
       "      <td>0.150656</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.003079</td>\n",
       "      <td>-0.162211</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.090225</td>\n",
       "      <td>-0.042135</td>\n",
       "      <td>-0.087721</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.154017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_type</th>\n",
       "      <td>0.013634</td>\n",
       "      <td>0.020200</td>\n",
       "      <td>-0.068732</td>\n",
       "      <td>-0.007774</td>\n",
       "      <td>-0.009998</td>\n",
       "      <td>0.003079</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.073051</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.015087</td>\n",
       "      <td>-0.776507</td>\n",
       "      <td>-0.068458</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.212893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fare</th>\n",
       "      <td>-0.014083</td>\n",
       "      <td>0.035106</td>\n",
       "      <td>0.917127</td>\n",
       "      <td>0.185992</td>\n",
       "      <td>-0.129692</td>\n",
       "      <td>-0.162211</td>\n",
       "      <td>-0.073051</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.091508</td>\n",
       "      <td>0.425216</td>\n",
       "      <td>0.395950</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.983444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>extra_fare</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tax</th>\n",
       "      <td>-0.009308</td>\n",
       "      <td>-0.022736</td>\n",
       "      <td>-0.080828</td>\n",
       "      <td>-0.501256</td>\n",
       "      <td>0.010869</td>\n",
       "      <td>0.090225</td>\n",
       "      <td>-0.015087</td>\n",
       "      <td>-0.091508</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.012988</td>\n",
       "      <td>-0.148891</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.089695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tip</th>\n",
       "      <td>-0.024436</td>\n",
       "      <td>0.008936</td>\n",
       "      <td>0.389773</td>\n",
       "      <td>0.083778</td>\n",
       "      <td>-0.028087</td>\n",
       "      <td>-0.042135</td>\n",
       "      <td>-0.776507</td>\n",
       "      <td>0.425216</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.012988</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.267483</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.555170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>toll</th>\n",
       "      <td>0.025840</td>\n",
       "      <td>0.003765</td>\n",
       "      <td>0.401863</td>\n",
       "      <td>0.246460</td>\n",
       "      <td>-0.153488</td>\n",
       "      <td>-0.087721</td>\n",
       "      <td>-0.068458</td>\n",
       "      <td>0.395950</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.148891</td>\n",
       "      <td>0.267483</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.403146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>surcharge</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_fare</th>\n",
       "      <td>-0.015078</td>\n",
       "      <td>0.033289</td>\n",
       "      <td>0.903529</td>\n",
       "      <td>0.184445</td>\n",
       "      <td>-0.127936</td>\n",
       "      <td>-0.154017</td>\n",
       "      <td>-0.212893</td>\n",
       "      <td>0.983444</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.089695</td>\n",
       "      <td>0.555170</td>\n",
       "      <td>0.403146</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 vendor  passengers  distance      rate  pick_location  \\\n",
       "vendor         1.000000    0.283619  0.015401  0.055897      -0.024097   \n",
       "passengers     0.283619    1.000000  0.033053  0.051624      -0.021166   \n",
       "distance       0.015401    0.033053  1.000000  0.119010      -0.119491   \n",
       "rate           0.055897    0.051624  0.119010  1.000000      -0.042557   \n",
       "pick_location -0.024097   -0.021166 -0.119491 -0.042557       1.000000   \n",
       "drop_location -0.005115    0.003783 -0.148011 -0.053956       0.150656   \n",
       "payment_type   0.013634    0.020200 -0.068732 -0.007774      -0.009998   \n",
       "fare          -0.014083    0.035106  0.917127  0.185992      -0.129692   \n",
       "extra_fare          NaN         NaN       NaN       NaN            NaN   \n",
       "tax           -0.009308   -0.022736 -0.080828 -0.501256       0.010869   \n",
       "tip           -0.024436    0.008936  0.389773  0.083778      -0.028087   \n",
       "toll           0.025840    0.003765  0.401863  0.246460      -0.153488   \n",
       "surcharge           NaN         NaN       NaN       NaN            NaN   \n",
       "total_fare    -0.015078    0.033289  0.903529  0.184445      -0.127936   \n",
       "\n",
       "               drop_location  payment_type      fare  extra_fare       tax  \\\n",
       "vendor             -0.005115      0.013634 -0.014083         NaN -0.009308   \n",
       "passengers          0.003783      0.020200  0.035106         NaN -0.022736   \n",
       "distance           -0.148011     -0.068732  0.917127         NaN -0.080828   \n",
       "rate               -0.053956     -0.007774  0.185992         NaN -0.501256   \n",
       "pick_location       0.150656     -0.009998 -0.129692         NaN  0.010869   \n",
       "drop_location       1.000000      0.003079 -0.162211         NaN  0.090225   \n",
       "payment_type        0.003079      1.000000 -0.073051         NaN -0.015087   \n",
       "fare               -0.162211     -0.073051  1.000000         NaN -0.091508   \n",
       "extra_fare               NaN           NaN       NaN         NaN       NaN   \n",
       "tax                 0.090225     -0.015087 -0.091508         NaN  1.000000   \n",
       "tip                -0.042135     -0.776507  0.425216         NaN  0.012988   \n",
       "toll               -0.087721     -0.068458  0.395950         NaN -0.148891   \n",
       "surcharge                NaN           NaN       NaN         NaN       NaN   \n",
       "total_fare         -0.154017     -0.212893  0.983444         NaN -0.089695   \n",
       "\n",
       "                    tip      toll  surcharge  total_fare  \n",
       "vendor        -0.024436  0.025840        NaN   -0.015078  \n",
       "passengers     0.008936  0.003765        NaN    0.033289  \n",
       "distance       0.389773  0.401863        NaN    0.903529  \n",
       "rate           0.083778  0.246460        NaN    0.184445  \n",
       "pick_location -0.028087 -0.153488        NaN   -0.127936  \n",
       "drop_location -0.042135 -0.087721        NaN   -0.154017  \n",
       "payment_type  -0.776507 -0.068458        NaN   -0.212893  \n",
       "fare           0.425216  0.395950        NaN    0.983444  \n",
       "extra_fare          NaN       NaN        NaN         NaN  \n",
       "tax            0.012988 -0.148891        NaN   -0.089695  \n",
       "tip            1.000000  0.267483        NaN    0.555170  \n",
       "toll           0.267483  1.000000        NaN    0.403146  \n",
       "surcharge           NaN       NaN        NaN         NaN  \n",
       "total_fare     0.555170  0.403146        NaN    1.000000  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr = df.corr(method ='spearman')\n",
    "corr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can drop features which show ``NaN`` correlation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(columns=['surcharge'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr = df.corr(method ='spearman')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Heatmap\n",
    "Completing the data science workflow from sourcing big data, wrangling it using Amazon Athena to well formed schema, bringing adequate sample data from Athena to notebook environment, conducting exploratory data analysis, and finally visualizing the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def heatmap(corr):\n",
    "    sns.set(style=\"white\")\n",
    "\n",
    "    # Generate a mask for the upper triangle\n",
    "    mask = np.zeros_like(corr, dtype=np.bool)\n",
    "    mask[np.triu_indices_from(mask)] = True\n",
    "\n",
    "    # Set up the matplotlib figure\n",
    "    f, ax = plt.subplots(figsize=(11, 9))\n",
    "\n",
    "    # Generate a custom diverging colormap\n",
    "    cmap = sns.diverging_palette(220, 10, as_cmap=True)\n",
    "\n",
    "    # Draw the heatmap with the mask and correct aspect ratio\n",
    "    sns.heatmap(corr, mask=mask, cmap=cmap, vmax=.3, center=0, annot=True, fmt=\"3.2f\",\n",
    "                square=True, linewidths=.5, cbar_kws={\"shrink\": .5})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x648 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "heatmap(corr)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_python3",
   "language": "python",
   "name": "conda_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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}