abcvoting: A Python package for approval-based multi-winner voting rules

Abstract

The Python package abcvoting is a research tool to explore and analyse approval-based committee (ABC) elections. First and foremost, it contains implementations of major ABC voting rules. These are voting rules that accept as input approval ballots, that is, the (binary) preferences of voters expressing which candidates they like or support. The output is a fixed-size subset of candidates, called a committee. Different ABC voting rules represent different approaches how such a committee should be formed. For example, there is a trade-off between selecting only widely supported candidates and choosing a committee that represent as many voters as possible. Much of the recent research has focussed on developing ABC voting rules that reflect the preferences of voters in a proportional fashion.

abcvoting is primarily intended for researchers interested in voting and related algorithmic challenges. The core content of abcvoting are implementations of a large number of ABC voting rules. These allow a user to quickly compute (and compare) winning committees for all implemented voting rules. In addition to computing winning committees, abcvoting can be used to verify axiomatic properties of committees. Axiomatic properties are mathematical formalizations of desirable features, e.g, fairness guarantees. Such properties are fundamental to the analysis and discussion of voting rules.