Identifying communities and ranking the drivers’ performance in Formula One

Abstract

This paper explores an approach for ranking the performance of Formula One drivers across different eras using network analysis methods, namely PageRank combined with community detection algorithms. Typically, attempts to rank driver performance have been limited, focusing solely on the evolution of technology and rules in the sport, and attempting to find the best individual driver, ignoring the fact that numerous good drivers never competed directly. To address these challenges, we collected data from all Formula One races since its inception in 1950 through 2022 and used network analysis to create communities of drivers based on their direct competition with each other. Within these communities, we then applied the PageRank algorithm to rank the drivers. Our approach has been shown to produce more meaningful and relevant driver rankings compared to the full graph PageRank results, and effectively recognises the dominance of drivers in their respective eras. Our methodology provides the basis for more sophisticated performance comparisons in sports with long histories and changing conditions.