This is the repository for Francisco Quiroga-Vergara's MSc project on Bayesian compositional forecasting.

Francisco Quiroga-Vergara, Eric Schulz, Nigel Harvey. "A Bayesian compositionality approach to judgmental forecasting."

In particular, it contains the reproducible code for the experiment and final thesis, as well as the raw data.

This is the html/js-implementation of the experiment. In particular,

• index.html: shows the informed consent for participants. This can be reused.
• experiment.html: shows the instructions of the experiments and the twelve plots (six per stage)
• source.js: functionality of experiment.html, i.e. data collection and delivery

This contains the collected data of our experiments:

• bayesian-forecasting-export.json: raw data of participants, as exported from Firebase
• analyze.py: the classes and functions to analyze the raw data

Currently, it only contains the proposal.