Does My Block Vote?

development

$ npm install
$ npm run watch

Generating Data

The voter turnout data comes from two sources

e.g. 2014 data:

Count of registered voters per district: http://www.elections.ny.gov/2014EnrollmentED.html
Number of votes per district: http://vote.nyc.ny.us/html/results/2014.shtml

Manual steps to setup data:

The enrollment data (# of registered voters) is given in the form of 5 pdfs (1 for each county)

To be able to use this data, we convert pdfs to xlsx (manually, using https://www.pdftoexcel.com/) And then to csv (by opening in excel as saving as csv - if prompted click "save active sheet" (the others are empty)) We now have 6 csvs that are converted to dataframes and cleaned up using data_build/gen_turnout_json.py

Original PDF files in: raw_data/2014_pdf

CSV files in: data/2014_csv

The votes per district is already a csv and just needs some parsing (also done using gen_turnout_json): data/votes_ed_level_2014

Once all csv files are ready, run gen_turnout_json to clean data, compute values, and write to json file.

To run gen_turnout_json:

Need all 5 enrollment csv files in data/2014_csv and votes csv file in data.

setup virtual env and activate:

$ python3 -m venv .venv
$ source .venv/bin/activate
$ pip install -r requirements.txt

$ python3 gen_turnout_json.py outputs a json file with data for each district to be used on front end / in map example:

"57023": {
    "voters": 1143,
    "elect_dist": "57023",
    "borough": "Brooklyn",
    "votes": 53,
    "percent": 5.0,
    "boro_rank": 1589,
    "rank": 4991,
    "grade": "F-",
    "color": "#f0f9e8",
    "2016_rank": 5018,
    "2016_percent": 39.0
}

and some overall_data:

"overall_data": {
    "avg_percent_2014": "23.2",
    "max_rank_2014": "4993",
    "avg_percent_2016": "59.6",
    "max_rank_2016": "5084",
    "by_borough_2014": {
        "Staten Island": {
            "max_rank": "265",
            "avg_percent": "30.4"
        },
        "Manhattan": {
            "max_rank": "1117",
            "avg_percent": "25.8"
        },
        "Queens": {
            "max_rank": "1159",
            "avg_percent": "22.7"
        },
        "Brooklyn": {
            "max_rank": "1590",
            "avg_percent": "22"
        },
        "Bronx": {
            "max_rank": "862",
            "avg_percent": "20.6"
        }
    }
}

The json file is saved in the data_build/parsed_data folder, and in src/static/data so that running the script updates both.
The script currently uses 2014 and 2016 data and so requires 2014 and 2016 data to be available, with 2016 data used only for text on the page (2016_rank, 2016_percent).

Generating Share Images

The share images are generated via gen_images.py in 4 steps

The folders are created (one for every district)
The map screenshots are created by visiting each url and cropping a screenshot and saved in a single folder for easier visual QA
Plots are created - used in share images and on district's page
Share images are generated and saved in district folder

Files used to generate the images, such as the emojis with grades live in image_generation_parts/

Generating the map legend

The legend is generated via: gen_legend.py, allowing colors to be easily changed

Neighborhood data

We overlay the neighborhood names using another dataset that has the central long and lat of each neighborhood source: https://data.cityofnewyork.us/City-Government/Neighborhood-Names-GIS/99bc-9p23/data
saved in raw_data/NHoodNameCentroids.csv
This data is converted to a json file: parsed_data/neighb_coords.js

Generating index.htmls

For each district, the index.html file varies slightly - these files are generated with gen_index_files.py.

The files differ for demo and prod so the script must be run twice with changes to .env.

For demo:

update .env to:

AWS_S3_KEY=voter-turnout
BASE_URL=https://demo.project.gothamist.com
IS_SCREENSHOTTING=false

Then run:

$ npm run build

In another window:

$ cd data_build
$ python3 gen_index_files.py
$ aws s3 cp ./district_folders_for_s3/ s3://gothamist-project-demo/voter-turnout --recursive --exclude .DS_Store --acl public-read --cache-control max-age=0 --content-type text/html

For Prod:

update .env to:

AWS_S3_KEY=voter-turnout
BASE_URL=https://project.gothamist.com
IS_SCREENSHOTTING=false

Then run:

$ npm run build

In another window:

$ cd data_build
$ python3 gen_index_files.py
$ aws s3 cp ./district_folders_for_s3/ s3://gothamist-project-prod/voter-turnout --recursive --exclude .DS_Store --acl public-read --cache-control max-age=0 --content-type text/html

Drawing the map

To draw the map, we use a geojson file exported from: https://data.cityofnewyork.us/City-Government/Election-Districts/h2n3-98hq.

To reduce filesize, we truncate longitude and latitude coordinates from 15 to 6 decimal places.

"the sixth decimal place in one decimal degree has 111,111/10^6 = about 1/9 meter = about 4 inches of precision"

Building

First make a copy of the the .env.sample file as .env:

$ cp .env.sample. env

These variables are used in the build process to prepare the static output for different environments.

Build with $ npm run build to output a build ready for an environment based on the values you use for the envvars below.

Key Name	Description
`AWS_S3_KEY`	The root-relative file path to assets. In development this can be left blank.
`BASE_URL`	The protocol and domain from which this app is served. Used for metadata only. In development this can be left blank.
`IS_SCREENSHOTTING`	Used to control some run-time settings when generating screenshots for share images. In development this can be left blank.
`GA_TRACKING_ID`	The GA ID for pageviews
`AWS_CLOUDFRONT`	Distribution ID. Used by Circle CI to invalidate the cloud front cache on deploy
`SIGNUP_ENDPOINT`	The zapier webhook endpoint to submit emails

nypublicradio / skunkworks_001