Software
Packages
NeuralNets4ElectionForensics
This repository contains Python scripts for the development of machine learning tools and models aimed at improving the accuracy and effectiveness of election forensics analysis.
Interactions
The package written in R facilitates estimation and visualization of marginal effects in complex scenarios by directly utilizing the variance-covariance matrix, rather than relying on the model object.
EFToolkit
The package written in R implements analysis of election fraud based on various election forensics methods.
CECscraper
The package written in R helps to scrape electoral data from the website of Central Election Commission of the Russian Federation. The package is no longer supported after 2021.
GPT3-Surveys
This repository contains scripts written in Python and data used for generation of synthetic responses of hard-to-reach elite members of the Russian elites based on the GPT-3 language model. For more information, see the paper: Kirill Kalinin(2022).Generation of Synthetic Responses to Survey Questions Using GPT-3: A Case of Hard-to-Reach Members of Russian Elites (based on the Survey of Russian Elites).
CLEA
The package written in R helps to generate party nationalization datasets for the CLEA project at three levels of aggregation: national, party and constituency levels.
Web Apps
Election Forensics Toolkit
The Election Forensics Toolkit website, developed by Walter Mebane and Kirill Kalinin, is a prototype that implements several methods that have been proposed as useful accuracy diagnostics. The website was developed as part of a project conducted by a team from the University of Michigan (Walter Mebane, Ken Kollman, Allen Hicken, Kalinin and Jonathan Wall) and the University of Maryland (David Backer), with funding from USAID.
Global Geo-Referenced Election Forensics Database
The goal of the Global Geo-Referenced Election Forensics Database project is to build a user-friendly interface convenient for exploration of geographic distribution of election fraud.