irs-migration

Visualizing taxpayer migration data from the IRS

The IRS publishes data on migration patterns within the US based on changes in taxpayers’ return addresses. This is a project centered on visualizing and exploring that data.

Initially just centering on San Francisco, I'm now working on turning the data-munging steps into an Express-based api to allow for exploration of any county in the US.

This means there are essentially two projects within this one:

• a backend api for delivering data
• a frontend visualization for consuming that data

data sources

https://www.irs.gov/statistics/soi-tax-stats-migration-data
https://www.irs.gov/statistics/soi-tax-stats-migration-data-2014-2015

https://www.census.gov/topics/population/migration/guidance/county-to-county-migration-flows.html https://www.census.gov/topics/population/migration/data/tables.html

census data

• 2010-2016
• 2000-2010

command line dependencies

• node
• npm
• yarn
• topojson
• geoprojection
• ndjson-cli
• shapefile
• imagemagick

Install missing command line dependencies by following directions on website and something like:

• downloading directly
• brew install {pkg}
• npm install --global {pkg}
• yarn global add {pkg}

to get all the data

1. clone this repo
2. npm install
3. npm run get-data
4. npm run atlas
5. npm run munge

api

to install

• get the data ^^
• irs-migration/munge/pg$ ./parse-county-pg.sh
• irs-migration/munge/pg$ node getseeddata.js
• irs-migration/api$ yarn install (yes, the backend uses yarn while the frontend uses npm.)
• make sure your db is running
• irs-migration/api$ ./node_modules/.bin/sequelize db:migrate
• irs-migration/api$ ./node_modules/.bin/sequelize db:seed:all

to develop

• irs-migration/api$ yarn run start:dev

stack

• express
• sequelize
• postgres

visualization

Currently, only focuses on San Francisco and consumes a static file created by npm run munge.

to develop

• get all the data ^^
• use webpack:
  • npm run start
• map code is in index.js

dependencies

• turf js
• d3
• webpack

to add support for a particular county

Say, Manhattan (FIPS code 36 061):

• get all the data ^^
• irs-migration$ ./munge/parse-county.sh 36 061
• files are now in data/36061:
  • inflow/outflow csvs
  • topojson of destination counties

TODO

• TODO in download.sh having to do with character encoding conversion
• change color scale to some sort of threshold scale (automatic)
• connect centroids with lines
• use circles on map instead of coloring counties
  • enter/exit morph between circle and county path
• bumps chart of top (10? 20?) counties year-to-year
• barchart of migration by state
  • sankey/sunburst/tree? (county-> state-> sf)
• select county to compare over time
• combine gif-munge and gif-create into single node script?
• note: 06075inflowcombined.csv has duplicate rows for cook county IL

Done

• command line generation of map
  • see munge/gif-munge.js and munge/gif-create.sh
• separate build dev/dist webpack functionality
• better mouseover tooltip
• zoom map
• line chart of total immigration/emigration year-to-year
• allow user to choose different year

FIPS ref

California 06

• 075 San Francisco
• 001 Alameda (Oakland, Berkeley, East Bay)
• 041 Marin
• 081 San Mateo (Silicon Valley)
• 097 Sonoma
• 055 Napa
• 013 Contra Costa
• 085 Santa Clara
• 087 Santa Cruz
• 095 Solano

New York 36

• 005 Bronx
• 047 Kings (Brooklyn)
• 061 New York (Manhattan)
• 081 Queens
• 085 Richmond (Staten Island)

Texas 48

• 201 Harris (Houston)
• 029 Bexar (San Antonio)
• 113 Dallas
• 453 Travis (Austin)
• 457 Tyler

Illinois 17

• 031 Cook (Chicago)
• 019 Champaign

Virginia 51

• 770 Roanoke City
• 161 Roanoke County
• 760 Richmond City
• 159 Richmond County

Washington, DC 11 001

gifs

To make gif of migration into New York County, NY (FIPS code 36061):

1. irs-migration$ ./munge/parse-county.sh 36 061
2. irs-migration$ cd munge
3. munge$ node gif-munge.js 36 061 in
4. munge$ ./gif-create.sh 36 061 in

To make legend for gifs

1. edit html file munge/legend.html
2. change variable vals to the contents of legendValsin.json/legendValsin.json
3. choose desired colorArray
4. open munge/legend.html in web browser
5. screenshot