top of page

Key Insights

from Data Analysis

District-level analysis of UK crime rates and house prices uncovering the relationship, its limits, and what really drives property values.

Top-level KPIs

📉

Overall Crime vs House Price

Weak overall relationship Pearson shows only a very small association

🔍

Crime Structure Insight

Crime type matters more than total crime volume

🚨

Strongest Local Effect

Anti-social behaviour matters most (r ≈ −0.32 within-district)

📊

PCA Result

Two crime patterns explain ~70% of crime variation

📈

Regression Insight

The full crime-type model achieved R² ≈ 0.40

Key Insights

Weak Overall Relationship

  • Total crime and house prices show only a weak overall correlation

  • Crime alone is not a strong general predictor of housing values

  • Aggregate crime statistics obscure the patterns that really matter

Short-Term vs Long-Term Effects

​​

  • Within-district analysis shows a stronger relationship (r ≈ −0.32) vs overall (r ≈ −0.03)

  • Recent local changes in crime have a greater impact than long-term averages

Perception vs Reality

  • Spearman shows anti-social behaviour has a stronger negative association (ρ ≈ −0.30) than most crime types

  • Visible disorder has a greater influence on housing demand than overall crime levels

  • Buyer perception of neighbourhood quality drives price sensitivity

Evidence of Urban Complexity

  • Theft from the person shows a positive correlation (r ≈ +0.37) with house prices

  • Bicycle theft also positive (r ≈ +0.28); both reflect urban concentration effects

  • High-value areas experience higher activity-related crime, not higher risk

Limitations of Single-Metric Analysis

  • Total crime shows a very weak relationship (r ≈ −0.03) with house prices

  • Crime-type analysis reveals stronger patterns up to |ρ| ≈ 0.30

  • Aggregated metrics hide meaningful relationships present at a more detailed level

Statistical Significance vs Practical Impact

  • Most results are statistically significant (p < 0.001) due to large sample size (N ≈ 34,000)

  • Even correlations as small as |r| < 0.10 appear significant at this scale

  • Statistical significance ≠ real-world impact; effect size must be interpreted carefully

 Final Conclusion

Total crime has only a weak relationship with house prices at the district level. However, certain disorder-related crimes, especially antisocial behaviour, have a more meaningful negative association with local housing values. Overall, housing prices respond more strongly to neighbourhood quality and crime composition than total crime volume.

Thank you 

  • LinkedIn
  • GitHub

© 2026 Dhruvbhai Jani · Final Year Project · Nottingham Trent University

bottom of page