README.md

---
title: Telemetry Analysis Dashboard
emoji: 📈
colorFrom: indigo
colorTo: gray
sdk: docker
pinned: false
tags:
  - telemetry
  - assetto-corsa
  - motorsport
  - data-visualization
  - dash
  - plotly
  - time-series
short_description: Assetto Corsa telemetry explorer dashboard.
---

# AC Telemetry Analysis — Demo Dashboard

Interactive dashboard for exploring Assetto Corsa telemetry sessions recorded by **[AC-Datalogger](https://github.com/nasim-raj-laskar/AC-Datalogger/tree/main)** — a high-frequency telemetry logger that reads directly from the game's Windows shared memory interface.

This Space ships with two pre-recorded demo sessions so you can explore the dashboard without running the logger yourself.

---

## What you're looking at

AC-Datalogger polls Assetto Corsa's shared memory (`acpmf_physics`, `acpmf_graphic`, `acpmf_static`) at ~65Hz and writes every sample to a session folder as Parquet + JSON. This dashboard loads those sessions and lets you explore 83 telemetry channels across five tabs:

| Tab | Contents |
|---|---|
| Driver Inputs | Speed, pedals, gear & RPM, steering angle |
| Vehicle Dynamics | G-forces, angular rates, wheel slip, wheel load |
| Tyres | Pressure, core temp, surface temp, wear, brake temps |
| Aero / Misc | Ride height, turbo boost, environment temps, fuel, driver aids |
| Track Map | `pos_x`/`pos_z` trace coloured by any channel, with a position scrubber |

The Track Map renders the full lap trace coloured by a selectable channel (speed, throttle, brake, lateral G, longitudinal G). The scrub slider moves the car marker along the trace entirely in the browser via `Plotly.restyle` — no server round-trip.

---

## Record your own sessions

The logger runs on Windows alongside Assetto Corsa. No plugins or game modification required — just enable shared memory in AC settings and run the script.

```bash
# install
pip install pandas pyarrow pyyaml

# record a session (while AC is running)
python main.py
```

Full setup, configuration, and captured feature reference → **[AC-Datalogger on GitHub](https://github.com/nasim-raj-laskar/AC-Datalogger/tree/main)**

---

## Session data format

Each recorded session is a folder named `track_car_YYYYMMDD_HHMMSS/` containing:

```
session_info.json      # car, track, max_rpm, max_fuel_kg, …
telemetry.parquet      # 83-column DataFrame at ~65Hz
telemetry.csv          # same data as CSV
```

`session_info.json` example:
```json
{
  "car": "ferrari_458_gt2",
  "track": "monza",
  "track_configuration": "gp",
  "max_rpm": 9000,
  "max_fuel_kg": 120.0
}
```

To use your own sessions with this dashboard, drop the session folder into `sessions/` and restart.