Congressional apportionment simulations for Columbia Journalism School's NewsCounts project, finding how state population swings (more specifically, census undercounts) could have gained/lost states additional seats

Analysis was conducted outside of markdown document due to running time, all source files included here.

This is a static simulation - it does not need to be rerun with updated data.

For the full analysis and summary: bwliv.github.io/apportionment

Guide to files below - when two files are noted, one uses 2010 population data, while the other uses 2018 American Community Survey estimates to roughly estimate 2020 population

statepops2010 + 2018ACS

aforementioned population data from 2010 and 2018 (which will be used as estimation for 2020)

apportioning + apportioning2018nums

source code for deterministic simulations generating estimates for changes state undercount needed to gain or lose house seats

seat_loss_gain + seat_loss_gain_2018

results from the above

calculatetotals + calculatetotals2020

generates summaries using these numbers, yielding the below pairs of tables

seatlosses + seatlosses2018

estimates for population (i.e. census count) change needed to gain/lose seat

seatlosses_per + seatlosses_per2018

same as above, but as percentage of state population

state_pers_within_uc + state_pers_within_uc2018

finds states that could have gained/lost a seat because of undercount of lack thereof

apportionment

markdown/project for these files

undercount

undercounting data for the 2010 census (see source files for original .pdf source)

archive_source_and_draft_files

used to generate data in main repo