Toolradar tool-page session snapshot by category
Extracted 5 November 2026, UTC. This read-only query outputs category aggregates only.

BEGIN READ ONLY;
COPY (
 SELECT c.slug AS category_slug, c.name AS category_name,
        COUNT(*) AS qualifying_tools,
        SUM(s.value)::bigint AS summed_tool_sessions,
        MIN(s."computedAt") AS snapshot_at
 FROM "ToolMetricSnapshot" s
 JOIN "Category" c ON c.id = s."categoryId"
 JOIN "Tool" t ON t.id = s."toolId"
 WHERE s.metric = 'views_30d' AND t.status = 'published' AND s.value IS NOT NULL
   AND s."computedAt" = (SELECT MAX("computedAt") FROM "ToolMetricSnapshot" WHERE metric = 'views_30d')
 GROUP BY c.id HAVING COUNT(*) >= 8
 ORDER BY summed_tool_sessions DESC, c.slug
 LIMIT 10
) TO STDOUT WITH CSV HEADER;
COMMIT;

Each underlying metric is the number of distinct Umami session IDs on an individual
/tool page in the preceding approximately 30-day window at the 5 November computation.
Only tool pages with at least 50 such sessions qualify for the underlying metric.
The snapshot applies an eight-tool minimum for the category cohort. A sum of per-tool
sessions is not a count of distinct sessions or people across a category: the same
session can visit multiple tools. A tool can belong to multiple categories, so rows
overlap. The source describes bots as excluded by analytics; the underlying SQL does
not apply an additional bot filter, so bot exclusion cannot be independently guaranteed.
A session does not establish that the visitor is a buyer, lead or purchaser.