TL;DR
This study investigates how sports clubs and events influence "expressive crimes" (like violence and public disorder) in London, using advanced spatial-temporal analysis. It finds that while sports clubs generally have a crime-reducing effect, major football matches can significantly increase these crimes in their immediate vicinity, driven by crowd dynamics and heightened emotions.
What Problem Does This Paper Address?
The relationship between sports and crime is complex and debated, with some studies suggesting sports can deter crime by promoting positive social behaviors and structured activities, while others highlight how large sporting events can act as "crime attractors" due to increased crowd density and emotional arousal. Existing research often relies on literature reviews or lacks empirical, data-driven findings that incorporate the vast potential of urban big data. This paper aims to bridge this gap by quantitatively mapping the spatial and temporal associations between sports clubs, major sporting events, and expressive crimes in London, providing evidence-based insights for urban crime prevention strategies.
Methods
The study employs a multi-stage analytical framework:
- Bivariate Moran's I Mapping: To visualize the spatial association between the density of sports clubs and expressive crime rates across London wards, identifying "hotspots" where they co-occur.
- Seasonal Trend Decomposition using Loess (STL): To analyze time-series data and compare crime patterns on match days versus non-match days, specifically around the Tottenham Hotspur Stadium. This helps to identify temporal associations and surges in crime.
- Difference-in-Differences (DiD) Approach: A quasi-experimental method to assess the causal effect of football match days on local crime dynamics. It compares crime trends in a "treated" ward (near the stadium) with "control" wards. An event-study framework was used to check the parallel trends assumption.
- Geographically and Temporally Weighted Regression (GTWR): To explore the driving mechanisms behind the space-time variation of expressive crimes. This advanced model allows regression coefficients to vary across both space and time, capturing dynamic and localized effects of factors like population mobility and sports club presence.
Datasets
- Crime Records: Point-level crime records across London (UK Police Open Data Portal, 2023) and daily ward-level crime counts in Haringey (Haringey Council, 2023).
- Sports Data: Number of sports clubs per ward in London (London Sport, 2023) and match schedules for Tottenham Hotspur Stadium (Premier League official website, 2023).
- Socio-demographic Covariates: UK Census 2021 data (demographic and socioeconomic indicators) and Index of Multiple Deprivation (IMD) 2019 data (composite deprivation score).
- Mobility Data: Daily underground passenger flow data (tap-in and tap-out counts) from Transport for London (TfL, 2023).
Key Insights and Ideas
- "Event-Driven Spaces": Areas around major sports stadiums, particularly during events, become "event-driven spaces" where increased crowd mobility and heightened emotional behaviors can trigger spikes in expressive crimes.
- Divergent Effects of Sports: While organized sports activities generally appear to suppress expressive crime by occupying leisure time and fostering social cohesion, their impact is not uniform and can vary significantly based on local socio-spatial conditions.
- Mobility as a Consistent Crime Driver: Surges in public transport usage (e.g., subway passenger flows) are consistently linked to increased expressive crime risk, acting as a stable crime-driving factor.
- Context Matters for Sports Clubs: In some low-density residential areas, sports clubs can be positively associated with expressive crime, suggesting that increased visitor inflows and social interactions in quieter areas can sometimes lead to more conflicts.
- Seasonal Influences: The crime-suppressing effect of sports clubs weakens during winter months and holidays, likely due to decreased outdoor activities and increased unstructured leisure time.
Key Results
- Hotspots Identified: Bivariate Moran's I mapping revealed hotspots of expressive crime and sports clubs clustered in central London and around major sports stadiums.
- Match Day Impact: Expressive crimes, particularly violent offenses, drug offenses, and public order violations, showed strong positive anomalies on Tottenham Hotspur match days. Property crimes, however, did not exhibit significant changes.
- Localized Impact: Time series analysis (STL) and DiD regression confirmed a strong, localized impact of match-day crowds around the stadium, with limited evidence of spillover into neighboring areas.
- GTWR Model Performance: The GTWR model achieved a high R² value of 0.83, indicating its effectiveness in explaining the spatial and temporal variation of expressive crimes.
- Mobility vs. Sports Clubs: GTWR analysis showed a consistent positive effect of subway passenger flows on expressive crimes, while sports clubs generally had a weaker and less consistent, often negative, effect.
Limitations and Future Work
- Data Availability Constraints: The study's case study was limited to Haringey borough and the Tottenham Hotspur Stadium due to data availability.
- Generalizability: Future research should extend these efforts to other stadiums across London to validate the generalizability of the findings.
- Advanced Modeling: Employing advanced machine learning and explainable AI methods could capture more complex nonlinear interactions between urban crimes and environmental variables, enhancing interpretability for crime prevention strategies.
- Broader Context: Further investigation into how major sporting events influence crime patterns in diverse urban contexts is recommended.
Summary updated