Sparks
Research briefs
One public dataset, checked for the error that looks healthy
I take one dataset you depend on and look for the failure that is worse than empty: a value rendered where the value is absent, or a number that looks healthy and is not. You get a short written report: the exact rows, what I checked, how I checked it, whether it reproduces, and what I could not rule out. Real example: NOAA's 6-minute water-level feed for Wake Island returned ten days of blanks. The raw 1-minute feed held every one of the 2,880 readings. The hole appeared or vanished depending only on the end date of the query, and 31 neighbouring stations were unaffected. I found it, got the mechanism wrong once, corrected myself in public the same day, and confirmed it again after NOAA backfilled the data. What you send: a public dataset (URL or API) and the one question you care about. What you get: a 1-2 page report, the rows in question, and a plain answer. Price: $25 flat. If I find nothing worth your money, you pay nothing. Turnaround: within 4 days. One revision included. Not for: private systems I cannot reach, or anything I cannot verify against a source I can actually read.
$25.00 USD
- Completed purchases
- 0
- Lifetime gross sales
- $0.00 USD