OpenAlex

Updated: Jan 16, 2026

Preprocessing

0. Install uv if needed and create the openalex uv environment by running uv init inside the openalex directory.
1. Download the OpenAlex snapshots from this link to a directory of your choosing (say, basedir).
2. To run the flattening script, first activate the uv openalex environment (if needed) by running source .venv/bin/activate inside the directory, then execute uv run preprocessing/flatten_openalex_files.py.
3. Open preprocessing/flatten_openalex_files.py and update the following variables: a. BASEDIR to the directory in Step 1. b. MONTH to the month of the snapshot, eg: may-2025
4. Scroll down to the if __name__ == '__main__': block near the end of the file.
5. Uncomment the lines one at a time and run the script flatten_<entity> functions to generate the flattened compressed CSV files. a. Start with flatten_merged_entries, then b. Then flatten_funders, flatten_concepts, ...., flatten_topics.
6. To flatten works, uncomment the block of code for flatten_works_v3. You flatten all JSONs at once, or do it N files at a time by changing the files_to_process variable to either all or an integer N.

Warnings:

• flattening authors and works take anywhere between 15 and 30 hours. The code will cache the files, so you should consider running it in batches by setting the files_to_process variable.