Broad coverage
Core categories across races, horses, sectionals, racecards, vet, weather, and pools.
Sha Tin and Happy Valley coverage with decades of race and off-track data for modelling, market analysis, and operations.
# Race Results — Happy Valley · 2025-07-16 (sample) Race_Key Horse_ID Rank Time Odds 20250716_839 HK_2022_H108 1 1:09.96 7.7 20250716_839 HK_2022_H297 2 1:09.97 6.2 20250716_839 HK_2022_H317 3 1:10.00 5.5 20250716_840 HK_2023_J421 1 1:09.28 3.6 ... # Full schema (40+ fields) shared after contact
0+
Years of history
0K+
Races recorded
0K+
Horses tracked
0.0M+
Structured records
Core categories across races, horses, sectionals, racecards, vet, weather, and pools.
Cross-source validation, point-in-time discipline, and post-race QC reduce research drift.
Historical bulk files, API access, and custom feeds aligned to your stack.
Dataset categories
Race date, venue, distance, class, and race-level metadata.
Pedigree, import type, ownership, and stable identifiers.
Finishing position, time, weight, connections, and closing odds.
Pre-race fields, recent form, rating moves, and priority signals.
Sectional positions, margins, and timings for pace research.
Injury, inspection, and clearance records for risk review.
Race-day weather and related environmental indicators.
Historical pool and odds-related data available on request.
Barrier trial text records for NLP and fitness workflows.
Historical race video assets delivered under contract scope.
Client-specific derived fields and dedicated feeds.
Bespoke datasets shaped around your modelling pipeline.
Full schema and sample data shared after contact.
"T+0 updates on race day.
— Horsorion data engineering standard
40 years of data, one key."
Integration-ready
Pull 40+ years of historical bulk straight from S3 (Parquet / CSV). Stream race-day updates with T+0 latency over a REST API. SFTP and VPC peering available on request.
# Pull a full 2024 season of race results (Parquet) $ aws s3 cp \ s3://horsorion-prod/results/year=2024/ \ ./local/ --recursive # One partition per year — incremental sync friendly download: results/year=2024/month=07/race_results.parquet download: results/year=2024/month=07/sectionals.parquet download: results/year=2024/month=07/odds_timeseries.parquet # IAM policy & bucket ACL configured per NDA
# GET race-day cards + odds time-series $ curl https://api.horsorion.com/v1/races \ -H "Authorization: Bearer $TOKEN" \ -G --data-urlencode "date=2025-07-16" \ --data-urlencode "venue=HV" 200 OK application/json { "race_key": "20250716_839", "venue": "HV", "runners": 12, "odds_snapshots": 47 } # API on waitlist · webhook push optional
How professional teams use Horsorion
"Build feature stores from historical sectionals and racecards to shorten research-to-production cycles."
"Needs dependable pre-race and post-race feeds for pricing models and risk monitoring."
"Consolidates vet, weather, and results data under one vendor to cut reconciliation overhead."
Quality assurance
01
Combine official and supplemental sources under one race key.
02
Multi-source checks and anomaly review reduce gaps and misalignment.
03
Preserve pre-race visibility to avoid look-ahead bias in backtests.
04
Post-race reconciliation keeps delivered datasets consistent.
05
Ship bulk files, API access, or custom feeds to match client workflows.
Integration Roadmap
Horsorion currently delivers via bulk files (CSV / Parquet / SQL) and secure transfer (SFTP / S3), minimising integration overhead. Real-time REST API and WebSocket are on the roadmap with an open waitlist.
CSV · Excel · Parquet · SQL dump · Avro. Delivered per season, 2–3 business days.
Long-term clients can access a read-only database query interface, secured by NDA and IP whitelist.
REST API and real-time odds WebSocket are in planning. API key + IP whitelist + TLS 1.3. Waitlist members notified before launch.
* Launch timing depends on client feedback and compliance review. No public commitment to specific dates.
Share your requirements and we will arrange the right dictionary, samples, and onboarding path.