Welcome to the code and data repository for our SIGIR 2022 perspective paper Reduce, Reuse, Recycle: Green Information Retrieval Research. This repository contains the raw data (see: experiments.toml) that we collected for each of the end-to-end ad-hoc retrieval pipeline (using the CodeCarbon library) and the additional emission data for household appliances and travel as seen in the paper. The code in this repository can be used to recreate all the tables and plots, and we also have the scripts used to capture the power and running times for the retrieval experiments.
python create_tables.py
- Creates both the main experiment table and the comparison table containing additional data.
- Outputs the two tables into the
output
folder. - This script generates a csv file for generating the two figures below.
python create_query_usage_plot.py
(creates Figure 1)python create_effectiveness_plot.py
(creates Figure 2)- Separately creates the two figures, and outputs them into the
output
folder.