StreamingBandit; Experimenting with Bandit Policies

TL;DR

StreamingBandit is a Python web app that facilitates the development, testing, and comparison of bandit policies in real-time field experiments, simplifying sequential treatment allocation.

Contribution

It introduces StreamingBandit, a novel tool that enables easy implementation and evaluation of bandit policies in applied social science studies.

Findings

Supports real-time policy testing and comparison

Enables quick development and reuse of bandit policies

Facilitates sequential treatment allocation in field experiments

Abstract

A large number of statistical decision problems in the social sciences and beyond can be framed as a (contextual) multi-armed bandit problem. However, it is notoriously hard to develop and evaluate policies that tackle these types of problem, and to use such policies in applied studies. To address this issue, this paper introduces StreamingBandit, a Python web application for developing and testing bandit policies in field studies. StreamingBandit can sequentially select treatments using (online) policies in real time. Once StreamingBandit is implemented in an applied context, different policies can be tested, altered, nested, and compared. StreamingBandit makes it easy to apply a multitude of bandit policies for sequential allocation in field experiments, and allows for the quick development and re-use of novel policies. In this article, we detail the implementation logic of StreamingBandit and provide several examples of its use.