CS Picks

How are rankings calculated?

CSPicks uses the Geometric Mean of adjusted publication counts across all research areas. This approach rewards institutions for having breadth across multiple areas of Computer Science rather than just dominating in one. The formula is: Score = (Π(adjusted_count + 1))^(1/num_areas).

What is "Adjusted Count"?

Each paper counts as 1.0 total. This value is divided equally among all co-authors. For example, if a paper has 4 authors, each author receives 0.25 credit. This prevents inflation of counts simply by having large author lists.

Where does the data come from?

Publication data is directly sourced from CSRankings. Historical affiliation data (tracking where a professor was in the past) is sourced from OpenAlex. Read Data & Acknowledgment for more info.

What is Historical Mode?

By default, rankings credit all of a professor's papers to their current institution. Historical Mode changes this attribution: it credits papers to the institution where the professor was affiliated at the time of publication. This provides a more accurate view of a department's output for a specific past time window.

What are the Conference Set options?

This filter determines which conferences are included in the rankings. Four options are available:

• CSRankings (All): Includes ALL conferences from csrankings.org, including both default venues and next-tier conferences (ASE, ISSTA, ICDE, PODS, HPCA, NDSS, EuroSys, Eurographics, FAST, USENIX ATC, ICFP, OOPSLA, KDD).
• CSRankings (Default): Includes only the default set of top-tier conferences from csrankings.org, excluding next-tier venues. This matches the original CSRankings behavior and is the default option.
• CORE A*: Only includes conferences rated as A* (top 4%) by the CORE Ranking 2023. This is the most restrictive option.
• CORE A: Includes conferences rated as A (top 13%) by the ICORE Ranking 2026. This includes 103 conferences across all CS areas.