spatial correlationrelationship between women’s public safety and infrastructure
intro: mapping safety through infrastructure
this project for cpln6800 at upenn explores the link between infrastructure and women’s public safety in new york city. the focus is on crime in public spaces, especially open areas and transportation hubs. in nyc, over 200,000 incidents of violence against women were reported in just two years, with 25.88% occurring in public places. a 2022 uk survey found that 32% of women feel unsafe in public at night, and women are 10% more likely than men to feel unsafe in the subway.
background: the power of infrastructure
james scott (1988) and michael mann (1984) described infrastructure as a key tool for organizing society. city planning often reflects gender inequalities, embedding power dynamics in everyday infrastructure. public spaces can reinforce these inequalities, making women face restrictions on time and movement. fear of violence limits women’s access to public areas, impacting their confidence and activity levels. women’s safety should be a priority in planning, with features like visibility, escape routes, and good lighting considered early in the design process.
public space design: making spaces safer
safe and comfortable public spaces for women should have:
- sufficient illumination: proper street lighting boosts safety and reduces fear, especially in dark or isolated areas.
- openness: clear views and spacious areas enhance visibility and security.
- access to public facilities: the freedom to use restrooms, public transit, walk alone, or bike safely.
- social interaction: access to activities within walking distance fosters community and a sense of safety.
analysis: correlating crime and infrastructure
the study considered factors like pavement conditions, broken streetlights, vacant buildings, and access to transportation. a link was found between these factors and crime rates, showing that poor infrastructure can increase the risk of violence. although correlation isn’t causation, proximity to these issues raises the chances of incidents. the study also looked at access to public services and resources for women.
kernel density estimation
the correlation between the number of crimes and infrastructure features
kde analysis: uncovering infrastructure issues
the kernel density estimation (kde) results show distinct patterns and problem areas for different aspects of urban infrastructure:
- vacant and insecure buildings: the kde analysis reveals a high concentration of vacant and insecure buildings in brooklyn. this suggests a critical area of concern for urban planners and policymakers, especially when addressing women’s safety and the overall quality of neighborhoods.
- streetlight conditions: the kde results for streetlights highlight a significant concentration of broken streetlights in manhattan. this emphasizes the need for consistent streetlight maintenance to ensure pedestrian safety and comfort, particularly for women, during nighttime.
- pavement quality: the analysis of pavement conditions across new york city indicates a moderate overall quality. this finding suggests room for improvement, as better pavement conditions would enhance safety and provide a more enjoyable walking experience for all residents.
identifying correlations: linking infrastructure and crime
correlations were found between urban infrastructure factors and crime rates. while correlation does not imply causation, the results suggest that proximity to damaged streetlights, subway stations, poorly maintained facilities, and low-quality pavement increases the likelihood of incidents, particularly affecting women. different algorithms may weigh these factors differently, but other socioeconomic elements also show significant associations with crime rates.
the k-nearest neighbors (k-nn) algorithm, a non-parametric method for classification and regression, was applied to provide a quantitative analysis of spatial relationships. in this study, k-nn was used to assess distances between various geographical features and reported crime locations. the mean average distances from features like vacant buildings, broken streetlights, poor pavements, women’s resources, subway stations, and facilities were calculated. correlation scatterplots further explore the spatial relationships between these features and the number of crimes, offering deeper insights into the study area.
modeling insights: predicting crime and infrastructure impact
statistical analysis helps identify key factors linked to violent incidents, though different methods offer varied perspectives. a linear regression model was used to predict crimes against women in public spaces, using predictors like distance to damaged streetlights, vacant buildings, subway stations, poor pavement, women’s resources, population, rent, house value, poverty rate, and offense types. results show that increased distance from damaged streetlights, facilities, poor pavement, and subway stations correlates with lower crime rates, while sex crimes have a strong positive association.
linear regression model
a random forest model further explored the influence of these predictors, highlighting key factors like poverty rate, total population, and median rent as major contributors to crime rates. the model’s importance values provide a nuanced view of each variable’s impact.
multiple analytical approaches reveal the interplay between urban infrastructure, socioeconomic factors, and women’s safety. this study aims to inform urban planning decisions that prioritize safer public spaces for women. as leslie kern noted, male-centric city planning has shaped urban design, underscoring the need for diverse perspectives. including women in planning processes can help build more inclusive, equitable, and safer cities.
correlation matrix
feature importance the increase in mean squared error in random forest
discussion: designing for inclusivity
while statistics help identify factors in violent incidents, human elements complicate criminology. spatial analysis reveals patterns but has its limits. this study aims to guide decision-makers and urban developers in creating safer, more inclusive spaces for women. as leslie kern noted, cities have been mostly designed by men, leading to a male-biased urban landscape. including diverse perspectives in urban planning can create safer, equitable spaces for everyone, promoting a fairer and more harmonious city.
copyright @liuhaobing, hit me to read the paper