Multi-Track Performance DNA Analyzer

Guidelines
Dashboard Visualization
Driver's Performance Analysis

Inspiration

Motorsport performance analysis often focuses solely on lap times, ignoring the nuanced patterns that define a driver's strengths and weaknesses. My goal was to create a tool that goes beyond raw numbers, generating a “performance DNA” for each driver to reveal their unique driving characteristics (strengths and weaknesses) across multiple tracks. By combining robust machine learning, deep data analysis, and interactive visualizations, I tried to empower drivers, teams, and coaches with actionable insights to improve training, revise strategy, and development.

What it does

The Multi-Track Performance DNA Analyzer generates comprehensive driver profiles that capture speed, consistency, adaptability, and sector-specific performance across tracks. Key features include:

Driver DNA Profiling & archetype classification (Speed Demons, Consistency Masters, Track Specialists, Balanced Racers)
Multi-track performance comparison & sector-level analysis
Interactive dashboards and heatmaps for easy visualization
Personalized training recommendations and actionable insights
Supports both end-users via GUI and developers via command-line tools

How we built it

I built the application using Python 3.13 with a modern dark-themed CustomTkinter GUI.

Data processing: Pandas, NumPy
Machine learning: Scikit-learn, PyTorch (clustering, PCA, time series analysis)
Visualizations: Matplotlib, Seaborn, Plotly
Packaging: PyInstaller, Inno Setup for Windows distribution

The system processes telemetry, lap, sector, and weather data to generate performance fingerprints and classify drivers into archetypes.

Challenges we ran into

Handling large telemetry datasets while keeping the application responsive and fast (~30s for full dataset)
Designing an intuitive visualization system that clearly conveys multidimensional performance metrics
Integrating machine learning models for clustering, PCA, and feature extraction in a GUI-friendly workflow
Balancing depth of analysis with accessibility for end-users without technical expertise
Enabling usage of zipped folders with data
Packaging the application into setup.exe file to enable simple installation while preserving the functionality of Python-based scripts and libraries

Accomplishments that we're proud of

Successfully analyzed 155 driver-track combinations across 6 tracks and 38 drivers
Identified 4 distinct driver archetypes and generated actionable training insights
Developed an interactive, professional GUI that makes complex analytics understandable at a glance
Created a fully self-contained Windows installer requiring no additional dependencies
Built a robust data processing and ML pipeline that scales to multiple tracks and drivers