Explanation of Process Used:

We integrated the Project Sidewalk Seattle Accessibility dataset with additional Points of Interest (POI) data from the Seattle Open Data portal, including libraries, parks, hospitals, and transit locations where pedestrian access is critical. All datasets were cleaned and standardized in Jupyter Notebook to ensure consistent geographic coordinates and key attributes such as barrier severity and permanence.

Using pandas, numpy, matplotlib, and seaborn, we explored patterns through bar charts, box plots, and distribution analyses. Spatial accessibility was assessed using scikit-learn’s BallTree, identifying barriers within a 200-meter walking radius to approximate real-world pedestrian access. Aggregating results by barrier type, neighborhood, time of day, and POI proximity enabled a targeted assessment of accessibility risk, revealing that many barriers are both frequent and moderately to highly severe.

Quality of Exploration of Data:

The analysis explores the data across multiple dimensions, including barrier frequency, severity, permanence, geography, and human usage patterns. Barrier data was examined both globally (through citywide distributions) and locally (using neighborhood and POI proximity). Sidewalk usage data was explored by age group, mode of movement, mobility usage, and time of day to contextualize who is affected by accessibility barriers. By combining infrastructure data with human activity data, the analysis enables us to understand pedestrian accessibility from a lens of equity. 

Quality of Metric Created:

We created a composite POI accessibility risk score to quantify how sidewalk barriers affect access to essential services. The metric integrates barrier count, severity, and permanence into a single interpretable score, which unlike using a simple count-based metric, allows us to prioritize barriers that are more permanent and severe (which reflects real pedestrian difficulty). 

Description of Metric:

For each POI, all sidewalk barriers within a 200-meter walking radius are identified. The risk score is calculated as: Barrier count * average severity * (1 + proportion permanent). 

Entities Ranked Based on Metric Created: 

Using the POI accessibility risk score, hospitals were ranked to identify which facilities face the greatest accessibility challenges in their surrounding pedestrian environment. This ranking highlights disparities in access to healthcare infrastructure, providing a clear and actionable outcome of the analysis.