Brian Fry

Through this research collaboration role with FBK, I supported the Digital Twin project for the city of Bologna, Italy. I cleaned and standardized data sourced from Bologna's extensive open data platform in order to construct metrics of neighborhood character related to topics of socioeconomic profile, mobility, traffic impact, and neighborhood amenities. Using these metrics, I employed machine learning methods k-means cluster analysis and principal components analysis to build a model which categorizes the city's neighborhoods. By highlighting the similarities and differences across the neighborhood clusters, this model gives policymakers an effective summary of the patterns of weaknesses and strengths of the city's neighborhoods to inform their their decision-making processes.

Carried out analysis of productivity data to inform operations strategy for managing maintenance fleets in eight of the largest urban bike sharing systems in the United States. Built interactive Tableau dashboards to provide stakeholders visibility on trends in KPIs and the various factors that could affect them. Created and managed data ingestion pipelines in Azure to ensure accurate and updated data in our analyses.

Ingested and analyzed data from various sources for quarterly and annual reports on hospital performance and quality of care. Designed monitoring dashboards and static visuals to illustrate improvements and differences in KPIs across stratifications of hospitals as part of our quarterly reporting on quality of care. Managed a bi-weekly data pipeline that used claims data to publicly publish clinicians’ active enrollment status and practice locations for beneficiary reference. Queried claims data weekly for COVID-related mortalities to aid in monitoring and forecasting of the pandemic.

Designed interactive visualizations of Twitter chatter data to explore patterns of forced migration in Iraq using Python, Javascript, and HTML.

Aided and instructed 150 students in writing and debugging intro level C++ projects. Corrected and graded 5 project submissions per week.