Oscar Piastri: Star in the Making? Part I

Introduction

The man himself

Every so often a new Formula 1 driver bursts onto the scene and captures the attention of fans and media. They garner lofty labels like “future champion”, “young gun”, or “budding star”. Twenty-two-year-old Oscar Piastri has emerged as the latest such figure, on the heels of an impressive rookie 2023 season, punctuated by standout performances in the Belgian, Japanese, and Qatar Grands Prix, the latter of which included a sprint race victory. Piastri has largely built on his impressive rookie campaign, recently notching his first Grand Prix victory in Hungary.

As a fan of sports like baseball, hockey and (American) football, sadly I’ve found F1 lacking from a statistical standpoint. Sure, we can easily enough observe simple “outcome” metrics like points, race wins, and pole positions, but the absence of underlying “process” statistics, such as car traits — downforce levels, tire temperatures, battery states, etc. — saddles the outside observer with relatively sparse samples of only somewhat informative data. Furthermore, the level of contextual details unavailable to the public renders comparative analysis even more challenging. For example, teams rarely divulge the impact of damage or suboptimal car setups, and we rarely gain insight into fuel levels, especially in practice and qualifying sessions.

So, does that mean we should throw up our hands and abandon empirical analysis? Of course not! Rather, it means we make the best out of what we do have available, avoiding broad, sweeping conclusions in most cases and acknowledging caveats and shortcomings.

To that end, I set out to answer a specific, simple question: how did Piastri’s rookie season compare to others in recent F1 history? To place this within the lexicon of the data and broader business worlds, this would amount to the business problem definition. To answer this question, I identified two metrics to calculate:

1. A rookie driver’s average qualifying gap to his teammate. As noted by the greatest current F1 journalist, Mark Hughes, a driver’s peaks often provide the most telling indications of his potential. What better way to assess peaks than to hone in on the most demanding moment of a Grand Prix weekend, when the fuel levels come down, the power levels come up, the grippiest tires get fitted, and drivers have to summon the absolute limits of their skills?
2. A rookie’s average position finish relative to their teammate. If the previous metric speaks to outright performance, this speaks to consistency, adaptability, and versatility. And while generally speaking, the best qualifiers turn in the best race results, plenty of exceptions to the rule exist. Just ask Jarno Trulli, who none other than Fernando Alonso himself anointed as his most difficult teammate in qualifying, only to finish his career with a single win and 11 podiums. Or even consider current (and three-time) champion Max Verstappen, who has 61 career wins on “only” 40 career poles. At the end of the day, points determine the standings, and races obviously determine the points.

This series will unfold in the following installments, which I hope will depict my general approach to framing and solving data problems. I also devised this project as a means to practice skills relevant to my job as a Data Engineer. I invite you to come along for the ride (pun very much intended) and learn whether the numbers indicate the young Australian is as impressive as the surface results suggest!

I. Introduction (today)

II. Data Requirements Gathering (coming next week)

III. Data Architecture Overview

IV. Data Modeling

V. Development

VI. Analysis

VII. Next Steps & Enhancements