CampaignFinances

Concluded research on the inequalities present in campaign fundraising for Massachusetts State Elections. Final Report.

Topics of Focus

Disparity Caused by Out of District Donations

Economic

Occupational

Racial

Scripts Cited in Final Report

Locater/legFinderGlobal.py: Geolocate and Identify donor's political district from address data.
Race/raceProbs.py: Generate racial probabilites for each donor based off of US census surname API.
Race/BISG.py: Improve racial probabilities for each donor by considering the demographics of user town or neighborhood.

Navigation

Data Files:

• Senate Full Contribution Data.xlsx: Comprehensive Donation Data for 27 State Senators with Competitive Elections from 2008-Present

• Data On Senate Districts.xlsx: Demographic Data on each of the above districts.

• senate2018.xlsx: All 2016/18 Cycle political donations to all senate contendees, regardless of competition.

• housejobs.xlsx: All donations to house contendees with listed jobs for Out of District Analysis by Socioeconomic Status

• mass_house_full_update.xlsx/mass_senate_full.xlsx: All house contendees and senate contendees donations on 2016/18 Cycle.

• houseCandidatesUpdated.xlsx: Metadata on the campaign status and demographics of a candidate and their district for the 2016/2018 Cycle.

• demographics.xlsx: Racial demographics of MA towns & large Boston neighborhoods for BISG.

Code and Created Files:

1. District Locating: Testing ratios of In District and Out of District Donations using Shapefile Data

• legFinder.py: Address Extraction and Normalization, Geolocation, District Identification from SENATE2012_POLY.shp
• orderedDistricts.csv: Output from legFinder.py
• legMapper.py: Insures successful shapefile conversion by visualizing target districts

2. Overview Statistics: Visualizing Discrepancies between District Donation Ratios and Demographics

• SenateData_RAnalysis.Rmd: Calculating In District Vs. Out of District Contribution Proportions

3. Contribution Analytics: In depth analytics of Individual Donations and Demographics

• completeDistributionAnalysis.m: More district location, preparing data for demographic-based analytics
• dataAnalysis.m: Linear Regressions investigating relation between donations and demographics

4. Wealth Distributions: Analysis of donations taking each individuals home price into account.

• donorHouseValuations.py: Using Home Valuation API to estimate home prices in Feeney's District
• FeeneyWithHomes.mat: MATLAB table combining Full Contribution Data with Home Estimates
• wealthDistributionAnalysis.m: Analyzing relations between home prices and donor behavior
• occupationsAndWealth.m: Running frequency analysis on donor's listed occupation
• AverageDonationsByProfession.twb: Tableau Workbook showing relation between donations and jobs
• fullOccupationAndWealth.m: Drawing relationships from Job titles
• jobIdp.xlsx: Table of Job Title, In District Donation percentage, and other demographic stats

5. 2018 Elections Only: Normalizing Occupation Distributions by using only 2018 Data

• LegFinderB.py: Modified legFinder.py with similar functionality for 2018 dataset
• orderedDistrictsB.csv: output from LegFinderB.py
• OccupationSenateData.png: table of average donations by job category

6. Locater

• legFinderGlobal.py: Modified from LegFinder.py to map districts of any state given a house or senate shapefile
• mass_house_full_update/massHouseFullFormatted.xlsx: Partitions of all donations to house races in 2016-2018 cycle = mass_senate_full/massSenateFullFormatted.xlsx: Partitions of all donations to senate races in 2016-2018 cycle = naiveDistrictAssigner.py: Assigning assumed district of recepient by identifying the most common district of origin of their donors. NAIVE... REQUIRES A MANUAL CHECK THROUGH MA's ONLINE REPRESENTATIVE LOOKUP TOOL
• csvWriter.py: writes legFinder and districtAssigner results to a csv

7. Race

• raceProbs.py: Scraping Census API to extract likelihoods of each donor's race
• raceWriter.py: Writing race proportions to races.csv
• raceAnalysis: Using CLT to estimate donor trends by race... E.g. Average Amount, Out of District Influence, etc...

8. Wealth Test

• randomValuationTest.py: Taking 1000 random samples from dataset and finding home zestimate
• randomValuationAnalysis.py: Exploring relationships between home price, amount, and district status

9. Multiple Donors

• candSummary.py: Exploring relationships between district demographic data and donation trends.
• bipartateGraph.py: Creating network of candidates connected by shared donors for Senate and House campaigns.

Tools and Attributions

- Python - R - MATLAB - Tableau -

• OCPF of Massachusetts Open Data
• United States Census
• OpenStreetMap-Nominatim
• Zillow Home Valuation API

Contributors

Matthew Katz - Boston College - katzmn@bc.edu

Justin Miller - Northeastern University - miller.justi@northeastern.edu

Sethu Odayappan - Harvard College - sodayappan@college.harvard.edu