Multiple linear regression of metro ridership in the context of football events: a case study of Metro de Lisboa

Special events, such as sports games, concerts, and festivals, pose significant challenges for efficient operational management, leading to disruptions in public transport systems, particularly urban metros. While existing studies have focused on predicting metro passenger flows during sport events based on weather or opponent type variables, little is known about the impacts of competition type, team form, position in the competition, or game results on both game attendance and metro demand. This study examined inbound and outbound passenger flows at the station-level in the Lisbon metro during football games. The database comprised 186 events spanning from 2017 to 2022, including 15-minute mobility data from three metro stations serving two stadiums in the city. Multiple linear regression (MLR) was applied to estimate attendance games. Results revealed statistical differences in game attendance based on competition and opponent types, as well as team situation in the competition. Also, there was an increase in metro demand up to 3,600 passengers on event days compared to non-event days. The MLR models incorporating all the variables retained significantly more data information than other model variants for both teams. Some significant attendance model predictors were related to opponent and competition types, and team’s situation before the games.