Similarity matrix
Which conferences draw the same companies, how those crowds shift from year to year, and which are growing. Use it to find conferences like the ones you already sponsor or speak at, and to see where the companies are heading. The many small regional PGDays count as one row, so they don't outweigh single large events.
Conference clusters, year by year
Each cell is a pair of conferences, darker when they share more companies. Similarity is cosine over the companies at each, with companies that appear almost everywhere (AWS, Google, Microsoft) counting less than niche ones, so a shared niche company is a stronger signal. Pick a year to cluster only that year's lists: conferences need at least 5 companies that year, and events that haven't happened yet don't count. Rows are ordered by hierarchical clustering, with the tree on the left and the clusters outlined on the diagonal, each named after its most central conference; the strip at the top is each conference's circuit, and the strip beside the rows is where its events happen. Hover a cell for counts; click to open the conference.
Loading…
Which clusters are growing, 2025 → 2026
For each cluster, its conferences that published sponsor lists in both years are compared like for like: how many distinct sponsor companies they drew in 2026 against 2025. 2026 counts only events held so far. Clusters with fewer than 3 comparable conferences are left out.
- Like AWS re:Invent▲ +236%
230 → 772 sponsor companies across 5 conferences compared: Google Cloud Next 20→296, AI Engineer 137→385, MCP Dev Summit / AGNTCon + MCPCon 19→166, DASH 31→28…
- Like IEEE ICDE▲ +35%
195 → 263 sponsor companies across 10 conferences compared: ICML 66→99, KDD 29→43, CVPR 41→54, MLSys 20→32…
- Like PGConf NYC▲ +20%
117 → 140 sponsor companies across 6 conferences compared: PGDays & regional Postgres events 90→114, Postgres Conference 12→17, PGConf NYC 20→22, PostgreSQL Conference Europe 29→31…
- Like Data Platform Next Step · Europe▲ +12%
25 → 28 sponsor companies across 3 conferences compared: Fabric February 13→18, Data Platform Next Step 9→9, Power BI Next Step 8→6
- Like Gartner Data & Analytics Summit▲ +6%
810 → 860 sponsor companies across 15 conferences compared: Gartner Data & Analytics Summit 29→144, Data + AI Summit 177→234, Data Decoded 33→72, Coalesce 26→40…
- Like EuroPython▼ −17%
486 → 403 sponsor companies across 4 conferences compared: EuroPython 29→31, SciPy Conference 19→14, PyCon 353→317, PyData 128→74
- Like Iceberg Summit▼ −18%
121 → 99 sponsor companies across 6 conferences compared: Airflow Summit 10→21, Iceberg Summit 23→33, Data Streaming Summit 22→24, Beam Summit 8→7…
- Like DataGrillen · Europe▼ −18%
98 → 80 sponsor companies across 5 conferences compared: SQL Konferenz 6→20, Data Saturdays 16→19, PASS Data Community Summit 28→20, SQLBits 24→14…
- Like OmniSci Converge▼ −29%
274 → 195 sponsor companies across 4 conferences compared: STAC Summit 80→95, NVIDIA GTC 72→82, World AI Cannes Festival 14→9, The AI Summit 144→37
Not enough comparable conferences: Like Community Over Code; Like ADE Summit · North America; Like DATACON Seattle; Like DACH AI, Data Analytics & Insights Summit · Europe; Like MLOps World · North America; Like Trino Fest; Like Data Day Texas · North America; Like Microsoft AI Tour; Like Arc of AI Conference · North America; Like MySQL BR Conf; Like EARL; Like DuckCon; Like JupyterCon · Europe; Like Neudata Data Summits; Like Minds Mastering Machines · Europe.
Circuit by circuit
Read across a row: each cell is the share of the row circuit's sponsor companies that were also present (as sponsors) in the column circuit. The diagonal shows how many sponsor companies each circuit has. The matrix is not symmetric: a small circuit can sit mostly inside a big one. The scale tops out at 40%; hover a cell for counts.
Scroll sideways to see the whole matrix.