Description

This repository contains the reproduction code for "User and Recommender Behavior Over Time" paper at FairUMAP 2025

Samira Vaez Barenji, Sushobhan Parajuli, Michael D. Ekstrand. 2025. User and Recommender Behavior Over Time: Contextualizing Activity, Effectiveness, Diversity, and Fairness in Book Recommendation. In Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization (UMAP Adjunct ’25). ACM. DOI: 10.1145/3708319.3733710. arXiv: 2505.04518 [cs.IR].

Environment Setup

We recommend using pixi to reproduce the environment.

1. Install pixi (if you haven’t already):

   curl -fsSL https://pixi.sh/install.sh | sh

2. Run a shell in the pixi environment:

   pixi shell

This will activate the environment and install all the required dependencies listed in pixi.toml.

Instructions

1. Clone the required data tools:

   Check out the PIReT/INERTIAL Book Data Tools into a directory named bookdata, placed alongside this repository:

   parent-directory/
   ├── bookdata/          
   └── goodreads-time/   
   

   Follow the instructions in the PIReT/INERTIAL Book Data Tools repository to generate your copy of the UCSD Book Graph.

2. Run the experiment pipeline:

   From inside the goodreads-time directory, run:

   dvc repro

   This will run the recommendation and evaluation pipeline.

3. Generate figures:

   Use the provided Jupyter notebooks to create the charts and figures from the experiment results.