What if racing legends met on the same track across the time?

F1 Legendary Drivers Performance Analysis System

INFO-I590: Usable AI (2025)

Group: Leo & Zic

This project is a performance analysis system for F1 (Formula One) legendary drivers based on historical data. The system analyzes race data from the Monza circuit between 1985-2024, using machine learning methods to evaluate and rank legendary drivers from different eras.

Core Functionality Implementation | 核心功能实现


1. Data Collection and Preprocessing | 数据采集与预处理


  • Retrieve historical race data through Ergast F1 API
    通过Ergast F1 API获取历史比赛数据

  • Process two different data formats before and after 1995

    分别处理1995年前后两种不同格式的比赛数据

  • Data cleaning and standardization processing

    数据清洗和标准化处理

2. Feature Engineering | 特征工程

  • Average Z-Score | 平均Z-Score

  • Z-Score variance | Z-Score方差

  • Best/Worst lap times | 最佳/最差圈速

  • Median performance | 中位数表现

  • Lap time decay rate | 圈速衰减率

  • Outlier ratio | 异常值比例

3. Performance Modeling | 性能建模


Using elastic net regression model, combining the following features | 使用弹性网络回归模型，结合以下特征：

  • Lap time statistical features | 圈速统计特征

  • Completion rate | 完赛率

  • Historical performance | 历史表现

  • Season factors | 赛季因素

4. Ranking System | 排名系统

Comprehensive evaluation based on | 基于以下因素综合评估：

  • Model prediction scores | 模型预测得分

  • Historical results | 历史成绩

  • Cross-era performance calibration | 跨时代表现校准

Check our

Github