About

A race companion, built in the open.

Rees Performance is a Formula 1 companion you open while watching — the schedule, the circuit, the stakes, the standings. Built by one engineer, in public.

I'm Alphonso Woodbury — a senior software and data engineer. This is where I point that at the sport I can't stop watching: a tool I actually use on race weekends, built to be fast, clear, and genuinely useful.

Rees is a family name — and, as it happens, one with a little F1 in it already. The site doubles as an open record of how I build: data engineering, AI, and design, all in one place.

How it's built · RPX

Six layers, one pipeline.

Data flows from source to screen. Each layer does one job and hands off cleanly.

01 Sources
Where the data comes from

Historical results, deep lap telemetry, and live session feeds — all from open F1 data.

Jolpica · results FastF1 · telemetry OpenF1 · live Weather
02 Pipelines
Ingest, validate, transform

Raw to clean to analysis-ready, in stages. Tested for freshness and integrity on every run.

dltdbtDagsterbronze · silver · gold
03 Storage
Raw, relational, time-series, vector

Columnar parquet for scale, SQL for serving, time-series for telemetry, vectors for search.

R2 · parquetDuckDBD1Vectorize
04 AI & Analytics
Predictions, queries, rules

A model for race forecasts, natural-language questions over the data, and a deterministic rules engine for answers that are exact and traceable.

PredictionsRAGWorkers AIRules engine
05 API
Serve it fast, push it live

Cached JSON at the edge for everything static; WebSockets for anything that moves during a session.

WorkersRESTWebSocketEdge cache
06 Interface
The companion you're using

Static-fast by default, with interactivity and visualization layered in only where it earns its place.

Static HTML/CSS3D tracksAstroD3
Across every layer ObservabilityData quality SecretsCI/CDInfra as code

Live today: the dataset, this companion, and the 3D Track Explorer. The rest is the build plan — shipped in tiers.

Sources

Real data, all the way to live.

SourceWhat it givesReal-time
JolpicaResults, qualifying, schedules — 1950 to nowPost-session
FastF1Lap data, car telemetry, tyres, stintsPost-session
OpenF1Live telemetry, positions, intervals, weatherYes
WeatherTrackside conditions and forecastsYes

Historical depth is solved. Real-time is an engineering problem, not an access one.

Real-time

Live timing, during the race.

A live feed fans out to every open browser through one stateful object per session. (On the roadmap.)

Source
OpenF1 live
Ingest
Worker
Buffer
Queue
State
Durable Object
Push
WebSocket
Screen
Browser

Scope

Built in tiers.

Each tier ships on its own and stands up the next.

Tier 0 · live
Companion + dataset
Tier 1
Auto-refreshing data
Tier 2
Predictions
Tier 3
Live timing
Tier 4
AI platform

← Back to the race