mp2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# Mini Project 2\n",
    "## Advanced Geopandas: spatial join and geoprocessing\n",
    "\n",
    "**Names:** Enrique Sanchez, Parker Addison"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be exploring several common geospatial computations in this mini project.  Our goal is to practice different requests against a single layer, and then to join several layers using real data.  In this project we will demonstrate geopandas mastery and also gain experience discovering datasets and critically evaluating metadata.\n",
    "\n",
    "Find multiple datasets, and for each one explore its metadata.  In addition, provide a brief description of the dataset including the data-integrity characteristics we explored in class, as well as any data quality considerations—especially those which may impact geometric processing.\n",
    "\n",
    "In this project there is one warm-up problem, and one large, real-world problem that the City of San Diego is actively trying to address.  The best creative and rigorous solutions to this larger problem will be asked to demonstrate to the ScaleSD folks who are visiting to give an invited talk about smart cities."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "import geopandas as gpd\n",
    "import pandas as pd\n",
    "\n",
    "# add packages as needed\n",
    "import shapely\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as mpatches\n",
    "import matplotlib.lines as mlines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## Part I.\n",
    "\n",
    "Using an administrative boundaries map of California, find San Diego county and return its CRS, polygon coordinates, area (use sq. miles), bounding box (use latitude and logitude), and a list of neighboring counties.\n",
    "\n",
    "To be more general, create a function to return the above geometric characteristics and a list of neighbors of any chosen polygon in a given input dataframe.  The chosen polygon will have a given `name` which exists in a given `name_field` in the dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part I Dataset Description\n",
    "\n",
    "**Title:** CA Geographic Boundaries\n",
    "\n",
    "**URL:** https://data.ca.gov/dataset/ca-geographic-boundaries\n",
    "\n",
    "**Number of records:** 58\n",
    "\n",
    "**CRS:** EPSG: 3857\n",
    "\n",
    "**Description:**\n",
    "This data set contains the shapefile boundaries (EPSG:3857) for all 58 counties in California. It comes from the US Census Bureau's MAF/TIGER database and reflects the governmental units in California as of January 1, 2016. The data set was first released on February 2, 2017 and most recently modified on February 27, 2019. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "### YOUR CODE HERE ###\n",
    "\n",
    "def poly_info(input_gdf, name_field, name):\n",
    "    \n",
    "    sqfeet_to_sqmiles = 1 / 27878400\n",
    "    \n",
    "    # Note that it's likely only one entry will exist for a given name, but to\n",
    "    # be safe we should dissolve first.\n",
    "    gdf = input_gdf.dissolve(name_field).reset_index().copy()\n",
    "    \n",
    "    # Retrieving desired polygon\n",
    "    poly = gdf[gdf[name_field] == name].reset_index(drop=True)\n",
    "    \n",
    "    # Getting coordinate reference system\n",
    "    crs = poly.crs\n",
    "    \n",
    "    # Getting polygon coordinates\n",
    "    poly_coords = poly['geometry'].iloc[0]\n",
    "    \n",
    "    # Calculating area, ensuring epsg 2230 for accuracy\n",
    "    area = poly.to_crs({'init': 'epsg:2230'}).area.iloc[0] * sqfeet_to_sqmiles\n",
    "    \n",
    "    # Getting coordinates for bounding box\n",
    "    bounding_box = poly.to_crs({'init': 'epsg:4326'}).bounds\n",
    "    \n",
    "    # Getting neighbors of the polygon\n",
    "    neighbors_df = gdf[gdf.touches(poly_coords)]\n",
    "    neighbors = list(neighbors_df[name_field])\n",
    "    \n",
    "\n",
    "    return crs, poly_coords, area, bounding_box, neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Running the function with our real dataset.\n",
    "\n",
    "input_gdf = gpd.read_file('Counties/CA_Counties_TIGER2016.shp')\n",
    "name_field = 'NAMELSAD'\n",
    "name = 'San Diego County'\n",
    "\n",
    "crs,poly_coords,area,bbox,neighbors = poly_info(input_gdf, name_field, name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'init': 'epsg:3857'}\n",
      "<shapely.geometry.polygon.Polygon object at 0x7f085d8e4c50>\n",
      "4525.420823895433\n",
      "         minx       miny        maxx      maxy\n",
      "0 -117.611093  32.528837 -116.080952  33.50503\n",
      "['Imperial County', 'Orange County', 'Riverside County']\n"
     ]
    }
   ],
   "source": [
    "print(crs)\n",
    "print(repr(poly_coords))\n",
    "print(area)\n",
    "print(bbox)\n",
    "print(neighbors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Part II.\n",
    "\n",
    "In low and medium income neighborhoods (LMI), food choices are often limited, as these areas may lack healthy food stores or fresh food delivery mechanisms.  Or perhaps there are more fast food options than fresh produce stores.  Where the population is relatively poor, healthy food options isn't been an overwhelming priority for many households.  These areas are referred to as \"food deserts\".\n",
    "\n",
    "See these articles:\n",
    "* https://www.sandiegouniontribune.com/news/data-watch/sdut-san-diego-food-deserts-2016may15-htmlstory.html (see food desert definitions here)\n",
    "* https://www.npr.org/sections/thesalt/2013/03/13/174112591/how-to-find-a-food-desert-near-you\n",
    "* https://www.geospatial.institute/wp-content/uploads/2018/03/Mapping-Food-Desserts-in-Southern-California-Counties.pdf \n",
    "* https://www.sandiegoreader.com/news/2018/sep/04/stringers-imperial-beach-food-desert/# (a recent one, about IB)\n",
    "* https://www.kpbs.org/news/2018/mar/29/sdsu-students-take-food-deserts-national-city/\n",
    "* http://newscenter.sdsu.edu/sdsu_newscenter/news_story.aspx?sid=74890 \n",
    "* https://grist.org/food/americas-worst-food-deserts-map-lovers-edition/\n",
    "* https://ucsdcommunityhealth.org/news/promoting-produce-consumption/\n",
    "\n",
    "There was also a Master's thesis at UCSD: https://escholarship.org/uc/item/0s78j9g2\n",
    "\n",
    "Don't read them all, just skim to understand the issues and get ideas.\n",
    "\n",
    "We'll analyze the current business state of food deserts by looking at lists of businesses in San Diego as well as by using demographic census-tract data from SanDAG.\n",
    "\n",
    "We will use the City of San Diego tax certificates for our list of businesses.  This dataset is current and includes location, as well as detailed NAICS codes for each business.  We want to look at only active certificates, and we want to determine which types of businesses are relevant to healthy and unhealthy food options.  We have provided a list of healthy and unhealthy NAICS as part of the assignment.  You may use this or modify it if you wish.\n",
    "\n",
    "We will also use relevant spatial data layers from SanDAG.  Namely the Promise Zone and Census Tract boundaries.  Descriptions of promise zones can be found at the following links:\n",
    "* https://www.hud.gov/sites/documents/SAN-DIEGO_ZONE_3RD.PDF\n",
    "* https://www.sandiego.gov/economic-development/sdpromisezone/maps\n",
    "* https://www.sandiego.gov/economic-development/sdpromisezone\n",
    "\n",
    "Social-demographic data by census tract can likewise be downloaded from SanDAG.\n",
    "\n",
    "**Your tasks:**\n",
    "\n",
    "1. Download all relevant data reading them into geopandas dataframes and merging social-demographic information as necessary.\n",
    "\n",
    "2. Determine which businesses provide food options, and which should be considered healthy/unhealthy.  Justify your choices if you deviate from the list we provide.\n",
    "\n",
    "3. Determine which census tracts intersect with the Promise Zone, and which of the remaining tracts are within the boundary of the City of San Diego.\n",
    "\n",
    "4. For Promise Zone tracts, as well as other tracts within city limits, calculate the number of healthy and unhealthy food options within the tract.\n",
    "\n",
    "5. Compute the ratio of healthy to unhealthy options, and build a regression model to showcase possible dependence on social-demographic characteristics of the tract. Start with linear regression. Provide a brief description of regression results.\n",
    "\n",
    "6. Create a map displaying the above ratio in each census tract, and displaying the locations of healthy food options.\n",
    "\n",
    "We won't provide a skeleton for this problem. This is a much more open project than before. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 2 Dataset Descriptions\n",
    "\n",
    "---\n",
    "\n",
    "**Title:** Business Tax Certificates\n",
    "\n",
    "**URL:** https://data.sandiego.gov/datasets/business-listings/\n",
    "\n",
    "**Number of records:** 39978\n",
    "\n",
    "**CRS:** N/A\n",
    "\n",
    "**Description:** This data set includes all active businesses since 2008 in San Diego. It is maintained by the City of San Diego and has been frequently updated (last modified on April 18, 2019 although we are unsure of the version provided to us in this mini project).\n",
    "\n",
    "---\n",
    "\n",
    "**Title:** Promise Zone SD\n",
    "\n",
    "**URL:** http://rdw.sandag.org/Account/gisdtview?dir=Business\n",
    "\n",
    "**Number of records:** 1\n",
    "\n",
    "**CRS:** ESPG 2230\n",
    "\n",
    "**Description:** This data set contains the shapefiles for the San Diego Promise Zone and is current as of June of 2016. According to the metadata, the information is within a 5% range of accuracy and is provided for general information purposes only.\n",
    "\n",
    "---\n",
    "\n",
    "**Title:** Census Tracts 2010\n",
    "\n",
    "**URL:** http://rdw.sandag.org/Account/gisdtview?dir=Census\n",
    "\n",
    "**Number of records:** 627\n",
    "\n",
    "**CRS:** ESPG 2230\n",
    "\n",
    "**Description:** This data set contains the shapefiles for the 2010 census tracts for San Diego County. As can be expected, this data was collected for the purpose of the US census. The metadata suggests that the the positional accuracy of each tract (polygon) may not exactly match their exact locations.\n",
    "\n",
    "---\n",
    "\n",
    "**Title:** San Diego City Boundary\n",
    "\n",
    "**URL:** https://data.sandiego.gov/datasets/san-diego-boundary/\n",
    "\n",
    "**Number of records:** 70\n",
    "\n",
    "**CRS:** ESPG 2230\n",
    "\n",
    "**Description:** This data set contains the shapefiles for City of San Diego administrative boundary. It is maintained by SanGIS and has last been modified on April 9, 2019. The data is updated when notified by LAFCO about a new annexation or detachment in order to ensure the accuracy of the data set.\n",
    "\n",
    "---\n",
    "\n",
    "**Title:** Deomgrographic & Socio Economic Estimates\n",
    "\n",
    "**URL:** http://datasurfer.sandag.org/\n",
    "\n",
    "**Number of records:** 43890\n",
    "\n",
    "**CRS:** N/A\n",
    "\n",
    "**Description:** This data set by SANDAG contains demographic and socio economic estimates for households in San Diego based on the 2010 census. It is noted by SANDAG that the census did not include information about structure type or household income. Therefore, they acquired this information from other sources such as the County of San Diego. It is also noted that minor adjustments were made to more accurately reflect the population and housing distribution.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "### YOUR CODE HERE ###"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Download the data - promise zone, census tracts, jurisdictions (census tracts social-demographic data and businesses data are provided but you can download as well if you wish), read them into geopandas dataframes, and merge social-demographic data with tract boundaries as needed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "tracts = gpd.read_file(\"Census_Tracts/CENSUS_TRACTS_2010.shp\")\n",
    "# We noticed some rows that were exact duplicates of one another.  This could mess with our counts of healthy\n",
    "# and unhealthy food options, so let's keep just the first to be safe!\n",
    "#\n",
    "# For an example of duplicated rows, look at businesses where dba_name == \"ZANZIBAR CAFE III INC\".\n",
    "# This particular food option is located outside of San Diego City, so it wouldn't affect our final\n",
    "# analysis, but it doesn't hurt to be safe!\n",
    "businesses = pd.read_csv(\"sd_businesses_active_since08_datasd.csv\").drop_duplicates()\n",
    "socio = pd.read_excel(\"socio-demographic.xlsx\")\n",
    "\n",
    "sandiego = gpd.read_file(\"SanDiego/san_diego_boundary_datasd.shp\")\n",
    "promise = gpd.read_file(\"Promise_Zones/Promise_Zone_SD.shp\")\n",
    "\n",
    "# The socio-demographic data and the census tracts data have identical \"TRACT\" columns.\n",
    "#\n",
    "# The dataframe resulting from their merge should be converted into a GeoDataFrame in\n",
    "# order to utilize geopandas in future steps.\n",
    "#\n",
    "# By calling merge from the gdf, we preserve the crs.\n",
    "socio_geo = tracts.merge(socio, how=\"right\", on=\"TRACT\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2. Subset the businesses (figure out, from NAICS descriptions, which businesses are likely to provide healthy food options, and which unhealthy. You can use our suggestions above but also feel free to deviate from them as long as you justify your choices.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#! NOTE: We considered moving 'FULL-SERVICE RESTAURANTS' to the list of healthy NAICS, but ultimately\n",
    "#  decided against this decision when we found that places like McDonalds and Burger King were listed\n",
    "#  as full-service.\n",
    "#\n",
    "#  In a perfect world, we would have the time to create an additional classifier for full service\n",
    "#  restaurants to further distinguish between healthy and unhealthy restaurants.\n",
    "healthy_naics = [\"FOOD (HEALTH) SUPPLEMENT STORES\", \"GROCERY STORES\", \"GROCERY (EXCEPT CONVENIENCE) STORES\"]\n",
    "\n",
    "unhealthy_naics = [\"FOOD SERVICES & DRINKING PLACES\", \"LIMITED-SERVICE EATING PLACES\", \"FULL-SERVICE RESTAURANTS\",\n",
    "             \"LIMITED-SERVICE RESTAURANTS\", \"BEER, WINE & LIQUOR STORES\", \"TOBACCO STORES\", \"CONVENIENCE STORES\"]\n",
    "\n",
    "def add_points_from_lat_lon(df):\n",
    "    \"\"\"\n",
    "    Takes a pandas dataframe which has a 'latitude' and 'longitude' column and returns a GeoDataFrame with\n",
    "    the geometry column containing the corresponding Shapely Points.\n",
    "    \"\"\"\n",
    "    \n",
    "    points = df.apply(lambda row: shapely.geometry.Point(row.longitude, row.latitude), axis=1)\n",
    "    \n",
    "    # This will be in EPSG:4326 since we're using lon-lat points.\n",
    "    return gpd.GeoDataFrame(df, geometry=points, crs={\"init\": \"epsg:4326\"})\n",
    "\n",
    "healthy_businesses = add_points_from_lat_lon(\n",
    "    businesses[businesses.naics_description.isin(healthy_naics)]\n",
    ")\n",
    "unhealthy_businesses = add_points_from_lat_lon(\n",
    "    businesses[businesses.naics_description.isin(unhealthy_naics)]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3. Determine census tracts intersecting with the promise zone, and all other census tracts within the boundary of San Diego."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tracts *intersecting* with Promise Zone:\n",
      " [ 33.01  33.03  33.04  33.05  30.01  30.03  30.04  31.01  31.03  31.11\n",
      "  31.12  31.13  34.03  34.04  35.01  35.02  36.01  36.02  36.03  38.\n",
      "  39.01  39.02  40.    41.    46.    47.    48.    49.    50.    51.\n",
      "  52.    54.   118.02 119.02  99.02]\n",
      "Tracts *within* San Diego City boundaries:\n",
      " [ 15.    16.    17.    18.    19.    20.01  20.02  21.    22.01  22.02\n",
      "  23.01  23.02  24.01  24.02  25.01  25.02  26.01  26.02  27.02   1.\n",
      "   2.01   2.02   3.     4.     5.     6.     7.     8.     9.    10.\n",
      "  11.    12.    13.    14.    32.11  34.01  27.05  27.07  27.08  27.09\n",
      "  27.1   27.11  28.01  28.03  28.04  29.03  29.04  31.05  31.14  31.15\n",
      "  66.    68.01  68.02  69.    70.02  71.    73.02  74.    75.01  77.01\n",
      "  77.02  78.    79.03  42.    43.    44.    45.01  53.    55.    56.\n",
      "  57.    58.    59.    60.    61.    62.    63.    65.    83.29  83.3\n",
      "  83.31  83.33  83.35  83.36  83.37  83.4   83.41  83.43  83.44  79.07\n",
      "  79.08  80.02  83.01  83.03  83.05  83.06  83.07  83.1   83.11  83.13\n",
      "  83.27  85.03  85.04  85.05  85.06  85.07  85.09  85.1   85.11  85.12\n",
      "  85.13  83.45  83.46  83.47  83.48  83.49  83.5   83.51  83.52  83.53\n",
      "  83.54  83.55  83.56  83.57  83.58  83.59  83.61  83.62  83.63  83.64\n",
      "  83.65  83.66  85.01  85.02  95.02  95.06  95.07  95.09  95.1   95.11\n",
      "  96.02  96.03  96.04  97.03  97.04  97.05  86.    87.01  87.02  88.\n",
      "  89.01  89.02  90.    91.01  91.02  91.03  91.04  91.06  91.07  92.01\n",
      "  92.02  93.01  93.04  93.05  93.06 100.03 100.04 100.05 100.1  100.11\n",
      " 100.12 100.13 101.06 101.1  101.11 101.12 170.46 170.47 170.52 170.15\n",
      " 170.18 170.31 170.34 170.35 170.36 170.37 170.43 170.56 214.  ]\n"
     ]
    }
   ],
   "source": [
    "# Fortunately, `tracts`, `promise` and `sandiego` are all in EPSG:2230 already!\n",
    "#\n",
    "# QUESTION: Would geopandas take care of the conversion if two different crs's\n",
    "# were used?\n",
    "#\n",
    "# ANSWER: No.\n",
    "# ```\n",
    "# >>> gpd.sjoin(tracts, promise.to_crs(epsg=4326), how=\"left\", op=\"intersects\")\n",
    "# UserWarning: CRS of frames being joined does not match!\n",
    "# ```\n",
    "\n",
    "promise_tracts = gpd.sjoin(tracts, promise, how=\"inner\", op=\"intersects\")\n",
    "\n",
    "# Note that the boundaries for San Diego City have objectid 40.  We're interested in\n",
    "# \"other census tracts within the boundary of San Diego\", so we use \"within\" as the operation.\n",
    "sd_city_tracts = gpd.sjoin(tracts, sandiego[sandiego.objectid==40], how=\"inner\", op=\"within\")\n",
    "\n",
    "# We want to compare the Promise Zone to the rest of the city's tracts, not to themselves!\n",
    "#\n",
    "#! NOTE: Errors are being ignored because there are some tracts that *intersect* with the promise zone,\n",
    "#  but are not *within* San Diego City, so attempting to drop those indices causes the error.\n",
    "sd_city_tracts_not_promise = sd_city_tracts.drop(index=promise_tracts.index, axis=0, errors=\"ignore\")\n",
    "\n",
    "# If we want just the tract numbers, we can just access the TRACT column of `promise_tracts`\n",
    "# and `sd_city_tracts_not_promise`\n",
    "print(\"Tracts *intersecting* with Promise Zone:\\n\", promise_tracts.TRACT.values)\n",
    "print(\"Tracts *within* San Diego City boundaries:\\n\", sd_city_tracts_not_promise.TRACT.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4. For each census tract marked as belonging to the promise zone, and the other tracts within the city limits, calculate the number of healthy and not-healthy options within that tract."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Start by turning all of the lon-lat points into the same crs as our other geometries. (May take a while)\n",
    "healthy_businesses_2230 = healthy_businesses.to_crs({\"init\": \"epsg:2230\"})\n",
    "unhealthy_businesses_2230 = unhealthy_businesses.to_crs({\"init\": \"epsg:2230\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Let's find out which tracts our healthy and unhealthy food businesses are in.\n",
    "healthy_businesses_tracts = (\n",
    "    gpd.sjoin(healthy_businesses_2230, tracts[[\"TRACT\", \"geometry\"]], how=\"inner\", op=\"within\").TRACT\n",
    ")\n",
    "\n",
    "unhealthy_businesses_tracts = (\n",
    "    gpd.sjoin(unhealthy_businesses_2230, tracts[[\"TRACT\", \"geometry\"]], how=\"inner\", op=\"within\").TRACT\n",
    ")\n",
    "\n",
    "# We want to return two dataframes, indexed by tract, which contains columns counting the number\n",
    "# of healthy and unhealthy businesses within that tract, respectively.\n",
    "\n",
    "# Note that the value counts of `healthy(unhealthy)_business_tracts` gives the number of healthy(unhealthy)\n",
    "# businesses per tract (tract is the index).\n",
    "promise_health_counts = promise_tracts.TRACT.to_frame().set_index(\"TRACT\").assign(\n",
    "    healthy_count=healthy_businesses_tracts.value_counts(),\n",
    "    unhealthy_count=unhealthy_businesses_tracts.value_counts()\n",
    ")\n",
    "\n",
    "other_health_counts = sd_city_tracts_not_promise.TRACT.to_frame().set_index(\"TRACT\").assign(\n",
    "    healthy_count=healthy_businesses_tracts.value_counts(),\n",
    "    unhealthy_count=unhealthy_businesses_tracts.value_counts()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "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>healthy_count</th>\n",
       "      <th>unhealthy_count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TRACT</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>33.01</th>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33.03</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33.04</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33.05</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30.01</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       healthy_count  unhealthy_count\n",
       "TRACT                                \n",
       "33.01            NaN              3.0\n",
       "33.03            NaN              1.0\n",
       "33.04            1.0              4.0\n",
       "33.05            1.0              2.0\n",
       "30.01            NaN              NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "healthy_count      13\n",
      "unhealthy_count    25\n",
      "dtype: int64\n",
      "healthy_count       54\n",
      "unhealthy_count    139\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "display(promise_health_counts.head())\n",
    "print(promise_health_counts.count())\n",
    "print(other_health_counts.count())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finally we can merge these counts back with our tracts data.\n",
    "promise_tracts_health = promise_tracts.merge(promise_health_counts, on=\"TRACT\")\n",
    "other_tracts_health = sd_city_tracts_not_promise.merge(other_health_counts, on=\"TRACT\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "5. Compute the ratio of healthy and non-healthy options, and build a regression model showing how it depends on social-demographic characteristics, such as median income or number of people in lowest and highest income brackets. You may use sklearn on this dataframe by tracts to compute a linear regresson. Provide a brief description of regression results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We want:\n",
    "#\n",
    "# healthy | unhealthy\n",
    "# --------|----------\n",
    "#       i |      NaN   --> Max ratio\n",
    "#     NaN |        i   --> Zero\n",
    "#     NaN |      NaN   --> NaN (will be dropped)\n",
    "\n",
    "# There are some concerns that arise with this:\n",
    "# e.g. should a single healthy option without any unhealthy options truly warrant the maximum ratio?\n",
    "#\n",
    "# However, for the purpose of utilizing as much data as possible, it makes sense that we should find\n",
    "# a way to compute ratios where one of the counts is zero (NaN) as opposed to completely ignoring the\n",
    "# row!\n",
    "\n",
    "def compute_ratios(counts_df):\n",
    "    \"\"\"\n",
    "    Takes as input the healthy and unhealthy counts columns as a dataframe.\n",
    "    Returns the ratio of healthy/unhealthy with the additional conditions:\n",
    "    - non-null healthy counts divided by null unhealthy counts will return the maximum of all other\n",
    "      calculated ratios\n",
    "    - null healthy counts divided by non-null unhealthy counts will return a ratio of zero\n",
    "    \"\"\"\n",
    "    \n",
    "    normal_ratios = counts_df.healthy_count / counts_df.unhealthy_count\n",
    "    max_ratio = normal_ratios.max()\n",
    "    \n",
    "    def compute(row):\n",
    "        if pd.isnull(row.healthy_count) and pd.isnull(row.unhealthy_count):\n",
    "            return np.NaN\n",
    "        elif pd.notnull(row.healthy_count) and pd.isnull(row.unhealthy_count):\n",
    "            \n",
    "            #! IMPORTANT: THIS MAY COMMENTED OUT TO RETURN MAX OR RETURN NAN.\n",
    "            return max_ratio\n",
    "            # return np.NaN\n",
    "            \n",
    "        elif pd.isnull(row.healthy_count) and pd.notnull(row.unhealthy_count):\n",
    "            return 0\n",
    "        else:\n",
    "            return row.healthy_count / row.unhealthy_count\n",
    "    \n",
    "    return counts_df.apply(compute, axis=1)\n",
    "\n",
    "    \n",
    "promise_tracts_health[\"ratio\"] = compute_ratios(promise_tracts_health[[\"healthy_count\", \"unhealthy_count\"]])\n",
    "other_tracts_health[\"ratio\"] = compute_ratios(other_tracts_health[[\"healthy_count\", \"unhealthy_count\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of ratios for promise zone tracts: 26\n",
      "Number of ratios outside promise zone but within SD city: 144\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of ratios for promise zone tracts: %d\" % promise_tracts_health.ratio.count())\n",
    "print(\"Number of ratios outside promise zone but within SD city: %d\" % other_tracts_health.ratio.count())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average ratio for promise zone tracts: 0.233814\n",
      "Average ratio outside promise zone but within SD city: 0.176882\n"
     ]
    }
   ],
   "source": [
    "print(\"Average ratio for promise zone tracts: %f\" % promise_tracts_health.ratio.mean())\n",
    "print(\"Average ratio outside promise zone but within SD city: %f\" % other_tracts_health.ratio.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We're going to want to compare the ratio for a tract with various information about its income levels.\n",
    "\n",
    "# Before to can do this, we should cut down on the size of our socio_geo df by only looking at the most\n",
    "# recently reported statistics.\n",
    "socio_recent = socio_geo[socio_geo.YEAR == socio_geo.YEAR.max()]\n",
    "\n",
    "# Let's develop a couple functions to take this data (10 income bins and their corresponding counts)\n",
    "# and extract a single object of useful information.\n",
    "def extract_median_bin(df):\n",
    "    \"\"\"\n",
    "    Takes in a dataframe with INCOME GROUP and HOUSEHOLDS (a count for each bin) and returns the median\n",
    "    income group.\n",
    "    \"\"\"\n",
    "    \n",
    "    total_households = df.HOUSEHOLDS.sum()\n",
    "    \n",
    "    median_household_number = total_households / 2\n",
    "    \n",
    "    median_group = df.loc[df.HOUSEHOLDS.cumsum() >= median_household_number, \"INCOME GROUP\"].iloc[0]\n",
    "    \n",
    "    return median_group\n",
    "\n",
    "# The total number of households might be nice to look at.\n",
    "def total_households(df):\n",
    "    \"\"\"\n",
    "    Takes in a dataframe containing information for a single tract and returns the total number of \n",
    "    households within that tract.\n",
    "    \"\"\"\n",
    "    return df['HOUSEHOLDS'].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We can also return every piece of info as a series, so that we can add them as rows.\n",
    "def info_to_series(tract):\n",
    "\n",
    "    return tract[[\"INCOME GROUP\", \"HOUSEHOLDS\"]].set_index(\"INCOME GROUP\").HOUSEHOLDS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "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>TRACT</th>\n",
       "      <th>SHAPE_AREA</th>\n",
       "      <th>SHAPE_LEN</th>\n",
       "      <th>geometry</th>\n",
       "      <th>YEAR</th>\n",
       "      <th>ORDINAL</th>\n",
       "      <th>INCOME GROUP</th>\n",
       "      <th>HOUSEHOLDS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>Less than $15,000</td>\n",
       "      <td>552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>2</td>\n",
       "      <td>$15,000 to $29,999</td>\n",
       "      <td>565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>3</td>\n",
       "      <td>$30,000 to $44,999</td>\n",
       "      <td>504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>$45,000 to $59,999</td>\n",
       "      <td>328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>5</td>\n",
       "      <td>$60,000 to $74,999</td>\n",
       "      <td>197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>6</td>\n",
       "      <td>$75,000 to $99,999</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>7</td>\n",
       "      <td>$100,000 to $124,999</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>8</td>\n",
       "      <td>$125,000 to $149,999</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>9</td>\n",
       "      <td>$150,000 to $199,999</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>16.0</td>\n",
       "      <td>7.407379e+06</td>\n",
       "      <td>11329.61606</td>\n",
       "      <td>POLYGON ((6292472.284999996 1855719.495000005,...</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>$200,000 or more</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     TRACT    SHAPE_AREA    SHAPE_LEN  \\\n",
       "120   16.0  7.407379e+06  11329.61606   \n",
       "121   16.0  7.407379e+06  11329.61606   \n",
       "122   16.0  7.407379e+06  11329.61606   \n",
       "123   16.0  7.407379e+06  11329.61606   \n",
       "124   16.0  7.407379e+06  11329.61606   \n",
       "125   16.0  7.407379e+06  11329.61606   \n",
       "126   16.0  7.407379e+06  11329.61606   \n",
       "127   16.0  7.407379e+06  11329.61606   \n",
       "128   16.0  7.407379e+06  11329.61606   \n",
       "129   16.0  7.407379e+06  11329.61606   \n",
       "\n",
       "                                              geometry  YEAR  ORDINAL  \\\n",
       "120  POLYGON ((6292472.284999996 1855719.495000005,...  2016        1   \n",
       "121  POLYGON ((6292472.284999996 1855719.495000005,...  2016        2   \n",
       "122  POLYGON ((6292472.284999996 1855719.495000005,...  2016        3   \n",
       "123  POLYGON ((6292472.284999996 1855719.495000005,...  2016        4   \n",
       "124  POLYGON ((6292472.284999996 1855719.495000005,...  2016        5   \n",
       "125  POLYGON ((6292472.284999996 1855719.495000005,...  2016        6   \n",
       "126  POLYGON ((6292472.284999996 1855719.495000005,...  2016        7   \n",
       "127  POLYGON ((6292472.284999996 1855719.495000005,...  2016        8   \n",
       "128  POLYGON ((6292472.284999996 1855719.495000005,...  2016        9   \n",
       "129  POLYGON ((6292472.284999996 1855719.495000005,...  2016       10   \n",
       "\n",
       "             INCOME GROUP  HOUSEHOLDS  \n",
       "120     Less than $15,000         552  \n",
       "121    $15,000 to $29,999         565  \n",
       "122    $30,000 to $44,999         504  \n",
       "123    $45,000 to $59,999         328  \n",
       "124    $60,000 to $74,999         197  \n",
       "125    $75,000 to $99,999         101  \n",
       "126  $100,000 to $124,999          25  \n",
       "127  $125,000 to $149,999           0  \n",
       "128  $150,000 to $199,999           0  \n",
       "129      $200,000 or more           0  "
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "socio_recent[socio_recent.TRACT == 16]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'$30,000 to $44,999'"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "extract_median_bin(socio_recent[socio_recent.TRACT == 16])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Now we can just wrap that function with something that takes in a tract number and spits out useful\n",
    "# info from the socio dataframe.\n",
    "def get_tract_socio(tract):\n",
    "    \n",
    "    tract = socio_recent[socio_recent.TRACT == tract]\n",
    "    \n",
    "    return pd.concat([\n",
    "        pd.Series({\n",
    "            \"median_group\": extract_median_bin(tract),\n",
    "            \"num_households\": total_households(tract),\n",
    "        }),\n",
    "        info_to_series(tract)\n",
    "    ], )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We can now create some extra income features so we can make our model later on.\n",
    "\n",
    "# Unfortunately we can't do much with tracts that have no ratios so they will not\n",
    "# be included in our model.\n",
    "promise_with_ratios = promise_tracts_health[promise_tracts_health['ratio'].notnull()].fillna(0).copy()\n",
    "other_with_ratios = other_tracts_health[other_tracts_health['ratio'].notnull()].fillna(0).copy()\n",
    "\n",
    "# Let's now compute the median household income for every tract we do ratios for.\n",
    "promise_with_info = pd.concat([\n",
    "    promise_with_ratios, promise_with_ratios.TRACT.apply(get_tract_socio).set_index(promise_with_ratios.index)\n",
    "], axis=1)\n",
    "\n",
    "other_with_info = pd.concat([\n",
    "    other_with_ratios, other_with_ratios.TRACT.apply(get_tract_socio).set_index(other_with_ratios.index)\n",
    "], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f072f544fd0>"
      ]
     },
     "execution_count": 303,
     "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": [
    "plt.scatter(promise_with_info.ratio, promise_with_info.num_households)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 304,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f072f4aff28>"
      ]
     },
     "execution_count": 304,
     "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": [
    "plt.scatter(other_with_info.ratio, other_with_info.num_households)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Already we can see that concerns about heteroscedasticity may arise if we wish to use linear regression!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [],
   "source": [
    "income_group_to_value = {\n",
    "    \"Less than $15,000\": 7500,\n",
    "    \"$15,000 to $29,999\": 15000 + 7500,\n",
    "    \"$30,000 to $44,999\": 30000 + 7500,\n",
    "    \"$45,000 to $59,999\": 45000 + 7500,\n",
    "    \"$60,000 to $74,999\": 60000 + 7500,\n",
    "    \"$75,000 to $99,999\": 75000 + 12400,\n",
    "    \"$100,000 to $124,999\": 100000 + 12400,\n",
    "    \"$125,000 to $149,999\": 125000 + 12400,\n",
    "}\n",
    "\n",
    "order = [\n",
    "    \"Less than $15,000\",\n",
    "    \"$15,000 to $29,999\",\n",
    "    \"$30,000 to $44,999\",\n",
    "    \"$45,000 to $59,999\",\n",
    "    \"$60,000 to $74,999\",\n",
    "    \"$75,000 to $99,999\",\n",
    "    \"$100,000 to $124,999\",\n",
    "    \"$125,000 to $149,999\"\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating a dataframe of ratios, with additional info columns for:\n",
    "# - Corresponding tract\n",
    "# - Whether or not in Promise Zone\n",
    "# - Median income group for that tract\n",
    "# - Median income as an estimated value for that tract (simply the midpoint of the bin range)\n",
    "# - Number of households for that tract\n",
    "all_info = pd.DataFrame(\n",
    "    pd.concat([\n",
    "        promise_with_info.assign(promise=1),\n",
    "        other_with_info.assign(promise=0)\n",
    "    ], sort=False).drop([\"SHAPE_LEN\", \"geometry\", \"index_right\", \"Juris\", \"Name\", \"objectid\", \"code\"], axis=1)\n",
    ")\n",
    "\n",
    "all_info[\"median_income\"] = all_info.median_group.replace(income_group_to_value)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "promise\n",
       "0    0.176882\n",
       "1    0.233814\n",
       "Name: ratio, dtype: float64"
      ]
     },
     "execution_count": 335,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_info.groupby(\"promise\").ratio.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f072f7f24e0>"
      ]
     },
     "execution_count": 336,
     "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": [
    "all_info.plot(x=\"num_households\", y=\"ratio\", kind=\"scatter\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 677,
   "metadata": {},
   "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": [
    "all_info.groupby(\"median_group\").ratio.mean().reindex(order).plot(kind=\"bar\").set_title(\"Median Income Group\\nvs.\\nMean Healthy/Unhealthy Ratio\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 324,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f072f1f37f0>"
      ]
     },
     "execution_count": 324,
     "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": [
    "plt.scatter(all_info.median_income + np.random.normal(scale=2000, size=len(all_info)), all_info.ratio)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f072f2b6630>"
      ]
     },
     "execution_count": 325,
     "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": [
    "all_info.where(all_info.promise == 0).plot.scatter(\"median_income\", \"ratio\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 493,
   "metadata": {},
   "outputs": [],
   "source": [
    "promise_mask = all_info.promise == 1\n",
    "promise, other = all_info[promise_mask], all_info[~promise_mask]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 515,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor\n",
    "from sklearn.linear_model import LinearRegression, Ridge, Lasso\n",
    "from sklearn.preprocessing import OrdinalEncoder\n",
    "from sklearn.compose import ColumnTransformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 549,
   "metadata": {},
   "outputs": [],
   "source": [
    "possible_groups = socio_geo[\"INCOME GROUP\"].unique()\n",
    "possible_groups\n",
    "\n",
    "ordenc = OrdinalEncoder(categories=[possible_groups])\n",
    "# ordenc.fit(possible_groups.reshape(-1, 1))\n",
    "\n",
    "# y_train, y_test = ordenc.transform(y_train.to_frame()).ravel(), ordenc.transform(y_test.to_frame()).ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 672,
   "metadata": {},
   "outputs": [],
   "source": [
    "def income_to_ratio_regression(df, reg=LinearRegression()):\n",
    "    \"\"\"\n",
    "    Takes in a dataframe of info, either for all tracts both Promise Zone and Ohter, or just one category.\n",
    "    \n",
    "    Uses linear regression to predict the ratio for each tract, and produces a bar plot of the true\n",
    "    ratios versus the predicted ratios.\n",
    "    \"\"\"\n",
    "\n",
    "    X = df.drop([\"healthy_count\", \"unhealthy_count\", \"ratio\", \"TRACT\"], axis=1)\n",
    "    y = df[\"ratio\"]\n",
    "\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y)\n",
    "\n",
    "    pl = Pipeline([\n",
    "        (\"pre\", ColumnTransformer([\n",
    "            (\"groups\", ordenc, [\"median_group\"])\n",
    "        ], remainder=\"passthrough\")),\n",
    "        (\"reg\", reg)\n",
    "    ])\n",
    "\n",
    "    pl.fit(X_train, y_train)\n",
    "\n",
    "    y_test.sort_index().plot(kind=\"bar\", color=\"blue\", width=0.8, alpha=0.8)\n",
    "    plt.bar(y_test.sort_index().index, pl.predict(X_test), color=\"orange\", width=0.8, alpha=0.7)\n",
    "\n",
    "    return pl.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 673,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.042598147746999326"
      ]
     },
     "execution_count": 673,
     "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": [
    "income_to_ratio_regression(all_info)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 674,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.28328626679846436"
      ]
     },
     "execution_count": 674,
     "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": [
    "income_to_ratio_regression(promise)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 675,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.010021574074648099"
      ]
     },
     "execution_count": 675,
     "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": [
    "income_to_ratio_regression(other)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 676,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0009978637781826727"
      ]
     },
     "execution_count": 676,
     "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": [
    "# Let's try it with RandomForestRegressor instead.\n",
    "income_to_ratio_regression(all_info, RandomForestRegressor(100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 682,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.9912226572034679"
      ]
     },
     "execution_count": 682,
     "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": [
    "# Maybe the linear regression is getting confused by all of the zeros... what happens if we drop them?\n",
    "income_to_ratio_regression(all_info[all_info.ratio > 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 687,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-7.942851494419953"
      ]
     },
     "execution_count": 687,
     "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": [
    "income_to_ratio_regression(promise[promise.ratio > 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 688,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-14.061215593949678"
      ]
     },
     "execution_count": 688,
     "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": [
    "income_to_ratio_regression(other[other.ratio > 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 625,
   "metadata": {},
   "outputs": [],
   "source": [
    "# None of these are very good regressors! They are **very** far from being accurate nor precise."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 626,
   "metadata": {},
   "outputs": [],
   "source": [
    "# What if we try predicting median income group using the ratio?\n",
    "\n",
    "def ratio_to_income_classification(df):\n",
    "    \"\"\"\n",
    "    Takes in a dataframe of info, either for all tracts both Promise Zone and Ohter, or just one category.\n",
    "    \n",
    "    Uses random forest classification to predict median income group for each tract, and produces a scatter\n",
    "    plot of the true group versus the predicted group.\n",
    "    \"\"\"\n",
    "    \n",
    "    cols_to_delete = df.filter(regex=r\"median|\\$\").columns\n",
    "\n",
    "    X = df.drop([*cols_to_delete, \"TRACT\"], axis=1)\n",
    "    y = ordenc.fit_transform(df.median_group.to_frame()).ravel()\n",
    "\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y)\n",
    "\n",
    "    clf = RandomForestClassifier(100)\n",
    "\n",
    "    clf.fit(X_train, y_train)\n",
    "\n",
    "    return clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 627,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3023255813953488"
      ]
     },
     "execution_count": 627,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratio_to_income_classification(all_info)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 628,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.42857142857142855"
      ]
     },
     "execution_count": 628,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratio_to_income_classification(promise)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 629,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3055555555555556"
      ]
     },
     "execution_count": 629,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ratio_to_income_classification(other)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 633,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The classification of median income group based primarily on ratio is *decent* when you\n",
    "# consider that there are eight categories to choose from..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 2 Discussion of Linear Regression Results\n",
    "\n",
    "As we expected, it is challenging to create a very useful model from the data we currently have our hands on.  However, there are still trends in our data that are highlighted by the predictive models.\n",
    "\n",
    "There is indeed some relationship between income level counts in a tract and that tract's corresponding healthy/unhealthy ratio.  This can be seen in the exploratory histogram of \"Median Income Group vs Mean Healthy/Unhealthy Ratio\"—there is a spike in the ratio at the 100k-125k median income range (which is very high for a median!)\n",
    "\n",
    "First we tried using linear regression in order to predict the healthy/unhealthy ratio of a tract.  As inputs to this regression model, we fed in: all income group counts, the median income group, the total number of households, and whether or not the tract was in the Promise Zone.  This was very ineffective, and wildly inconsistent—often times predicting negative values or all zeros depending on the training set.  We had a theory that the massive volume of '0' ratios may have been negatively impacting the accuracy of the linear regression, so we tested our regression model with all the zeros removed, but we saw little improvement.\n",
    "\n",
    "We also tried using random forest regression instead of linear regression and noticed a moderate improvement, but still a huge amount of inconsistency.\n",
    "\n",
    "Next, we tried using random forest classification in order to predict the median income group of a tract.  This model resulted in an average test accuracy of approximately 33%, which seems bad, but is actually fairly good when you consider that there are eight income groups to choose from, and that we only scored on accuracy (which is a very harsh metric) when instead we should have taken the ordinality of the data into account.  Given more time, we would have liked to have computed a score that would take the distance between the predicted income level and the true income level into account, but we ran into issues while trying to get sklearn to use a different metric on our predictions.  However, this goes to show that there is definitely *some* merit to predicting a median income level for a census tract based on the healthy/unhealthy ratio.\n",
    "\n",
    "Perhaps one of the biggest issue we discovered while trying to run regression and classification was that the randomly sampled training set had a massive impact on the outcome of the model.  This is largely due to the small size of our dataset compared to the large degree of variation in the data—which was only accentuated when we passed in a partial dataset (e.g. only Promise Zone tracts) to the model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6. Generate a map showing the ratio of healthy and not healthy food options, and the locations of stores offering healthy food options. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [],
   "source": [
    "sdcity = sandiego[sandiego.objectid==40]\n",
    "\n",
    "sd_healthy_businesses = healthy_businesses_2230.where(\n",
    "    healthy_businesses_2230.within(sdcity.geometry.item())\n",
    ")\n",
    "\n",
    "sd_unhealthy_businesses = unhealthy_businesses_2230.where(\n",
    "    unhealthy_businesses_2230.within(sdcity.geometry.item())\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 664,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# %matplotlib notebook\n",
    "\n",
    "ax = sdcity.plot(figsize=(10,10), color=\"lightgray\", label=\"San Diego City\")\n",
    "\n",
    "promise_tracts.plot(ax=ax, color=\"darkgray\", label=\"Promise Zone\")\n",
    "sd_city_tracts_not_promise.plot(ax=ax, color=\"whitesmoke\", label=\"Other Tracts\")\n",
    "\n",
    "# We need to make sure that the promise tracts and other tracts are plotted on the same color range.\n",
    "#\n",
    "# We do this by setting vmin and vmax to the same values.  (Also this means we only need to create\n",
    "# the legend for one of the instances!)\n",
    "#\n",
    "# The vmin should be zero (if there are no healthy food options), and the vmax should be the maximum\n",
    "# ratio found.\n",
    "max_ratio = max(promise_tracts_health.ratio.max(), other_tracts_health.ratio.max())\n",
    "\n",
    "cmap = \"YlGn\"\n",
    "\n",
    "promise_tracts_health.dropna(subset=[\"ratio\"]).plot(\n",
    "    ax=ax, column=\"ratio\", cmap=cmap, vmin=0, vmax=max_ratio, edgecolor=\"dimgray\", linewidth=0.5,\n",
    "    label=\"Healthy/Unhealthy Ratio\", legend=True\n",
    ")\n",
    "other_tracts_health.dropna(subset=[\"ratio\"]).plot(\n",
    "    ax=ax, column=\"ratio\", cmap=cmap, vmin=0, vmax=max_ratio, edgecolor=\"lightgray\", linewidth=0.5\n",
    ")\n",
    "\n",
    "\n",
    "# There are a very large number of unhealthy options that are often closely grouped, the alpha is\n",
    "# set lower than that for healthy options for visual clarity.\n",
    "#\n",
    "# NOTE: The prompt asks for just the locations of health food options.\n",
    "msize=5\n",
    "sd_healthy_businesses.plot(ax=ax, color=\"limegreen\", alpha=1, markersize=msize+2, edgecolor=\"lime\")\n",
    "# sd_unhealthy_businesses.plot(ax=ax, color=\"r\", alpha=0.3, markersize=msize)\n",
    "\n",
    "\n",
    "# Tedious legend work.\n",
    "promise_patch = mpatches.Patch(color=\"gray\", label=\"Promise Zone\")\n",
    "other_patch = mpatches.Patch(color=\"whitesmoke\", label=\"Other Tracts\")\n",
    "healthy_patch = mlines.Line2D([], [], color=\"limegreen\", linestyle='', marker='o', label=\"Healthy Option\")\n",
    "unhealthy_patch = mlines.Line2D([], [], color='r', linestyle='', marker='o', label=\"Unhealthy Option\")\n",
    "ax.legend(handles=[promise_patch, other_patch, healthy_patch], bbox_to_anchor=(0, 1), loc=2)\n",
    "ax.set_ylabel(\"Healthy/Unhealthy Ratio\")\n",
    "ax.yaxis.set_label_position(\"right\")\n",
    "ax.tick_params(bottom=False, left=False, labelleft=False, labelbottom=False)\n",
    "\n",
    "ax.set_title(\n",
    "    \"Healthy/Unhealthy Ratio of Tracts Within and Outside of the Promize Zone, \\nand Location of Food Options\"\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "# Could you briefly present what you found (up to 5 mins) to start the discussion and get Greg's feedback?  You can demo your notebook, show some resultant graph, and talk through the results.\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "celltoolbar": "Edit 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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}