Work through RaceRove's roughly 18,000 races and comprehensively enrich races with missing or incomplete information.
"yes do this and also mark races as duplicates if there are duplicates on the racerove app"
RaceRove reports 18,075 active listings; its four public race sitemaps expose 13,750 canonical race URLs, and 4,325 listings are not exposed publicly at all. Cumulative page extraction covers 451 URLs — an inventory status, not a fully-enriched count. Conservative duplicate screening of all 13,750 public URLs surfaced 56 normalized-name candidate groups (116 pages); after comparing city, date, distances, organizer, registration and venue, exactly one group is confirmed as a duplicate.
Each lane is scaled to its own total. Hover a segment for exact counts. Inventoried, researched and fully enriched are distinct statuses — no race is yet fully enriched under the required-dimension rule.
One group, two RaceRove pages. Keep the record that carries the confirmed date; mark the other as duplicate. Preserve the duplicate's aliases and non-conflicting metadata during consolidation. Verified 2026-09-25.
All 114 remaining candidate pages across 55 normalized-name groups. Metadata comparison (city/state, date or usual month, distances, organizer, registration, venue) shows these are distinct events or lack evidence to merge.
| Group | Race name | City, state | Date / month | Distances | Status | Source |
|---|
The official event site and current third-party listings agree on the name, Hampton / Hampton Beach NH location, Half Marathon & 5K distances and the 2027-03-07 date. Prefer the official hamptonhalf.com rows over older third-party editions.
The ongoing model stores normalized tables: canonical metadata, registration cycles, five-year weather, official-vs-runner logistics, up to five sourced food places, RaceRaves synthesis, course/elevation sources and route points, discrepancies, unresolved-field research outcomes, and field-level provenance. A race counts fully enriched only when every required dimension is sourced or explicitly marked not_found_after_research — which is why 451 processed pages is an extraction milestone, not an enrichment count.
Method: RaceRove public race sitemaps (4 files) enumerated 2026-09-25, yielding 13,750 canonical race URLs of 18,075 reported active listings; 451 pages extracted to date. Duplicate screening normalized race names across all 13,750 URLs into 56 candidate groups (116 pages); one group confirmed by metadata comparison plus independent web corroboration (89 pages checked; 8 shown for space). Counts measure pages and URL groups, not events held. Candidate table shows a subset of columns; organizer, venue and registration fields were used in comparison but trimmed for width.