Hi r\/datasets<\/a> \u2014 disclosure first: I sell a paid version of these on Gumroad ($34, 83% off launch). I’m posting the free 20-record samples here because they’re genuinely useful on their own and the mod rules ask self-promotion to be labeled.<\/p>\n What’s in the free samples:<\/p>\n 20 niche datasets, each with 20 fully-enriched records as CSV + JSON. ~55,000 records total in the paid version (54,958 as of today). Topics:<\/p>\n Why I built them:<\/p>\n I kept needing structured, queryable lists of “all the X tools” for filterable directory builds. Awesome-lists and READMEs are great for browsing but useless for jq \/ SQL \/ search infrastructure. So I curate, normalize, validate (zero invalid records), enrich with stars\/downloads\/installs, and refresh.<\/p>\n Per-record fields are typed \u2014 categorizationTier rates each record 87-100% specific (vs vague “tool” labels). Open question for the sub: how do you handle tier-of-specificity in your own dataset categorization work? My current rubric is per-dataset config-driven but I’m curious what others do.<\/p>\n Free samples (CSV + JSON, MIT-style permissive): https:\/\/github.com\/futdevpro\/niche-datasets-free<\/a><\/p>\n Includes mega-sample.json (5 random records from each of the 20 datasets, 100 records total).<\/p>\n Paid version on Gumroad \u2014 $34 launch price (83% off $198 list), monthly refresh on AI Models Pricing because OpenRouter changes weekly, quarterly on the rest. Linked from the GitHub README if anyone wants the full thing.<\/p>\n Happy to answer questions about the catalog, methodology, or specific datasets.<\/p>\n<\/div>\n submitted by \/u\/Jhonny_Ronnie <\/a> \n
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
