GoKawiilICLR 2026 — Institutional Affiliations Dataset & Analysis
End-to-end pipeline that turns 5,356 ICLR 2026 accepted papers into a clean, PDF-derived institutional-affiliation dataset and a publication-ready treemap of who is shaping AI research right now.
This avoids the OpenReview-profile drift problem (where authors' current job appears on every paper they ever wrote — e.g. listing Wyoming as the affiliation for a paper actually written at UBC). Affiliations come from the paper's title block PDF, not from author profiles.
Follow me for more analysis like this, plus AI engineering & research insights: LinkedIn — linkedin.com/in/dmytrolopushanskyy
GitHub — github.com/DmytroLopushanskyy If this dataset or the pipeline is useful to your work, a follow / star is the easiest way to encourage me to keep publishing this kind of analysis.
The headline chart
Each rectangle is one institution sized by the number of accepted papers it appears on (counted once per paper, regardless of how many of the paper's authors are affiliated with it). Region cells are sized by the cumulative count of their top-50 institutions. Lighter shade = academia / research institute, darker shade = industry.
Square version (for social posts): charts/iclr2026_top50_treemap_unique_grouped_square.png
What's in data/
File What it is iclr2026_public.csv / .xlsx The main dataset. 5,356 accepted papers with PDF-derived authors and institutions, normalized institution canonical names, country/region, abstract, OpenReview URL. UTF-8 with BOM for Excel compatibility. iclr2026_institutions_ranked_unique.csv Top-N institutions ranked by unique-affiliation count (each institution +1 per paper). iclr2026_institutions_ranked_first_author.csv Same, but only counting the first author's institution. iclr2026_institutions_ranked_fractional.csv Same, with fractional 1/N credit per institution per paper. iclr2026_method_sensitivity.csv Side-by-side rank under all three counting methods, so you can see which institutions are robust and which are method artefacts.
... continue reading