Philly Vehicle Locator
The City of Philadelphia's Office of Fleet Management and Streets Department wanted increase the usefulness of GPS data transmitted from their vehicles in the field. We stepped in to help and developed what we call the Philly Vehicle Locator (PVL). PVL is a system that ingests streaming GPS data from all fleet vehicles, and processed them using a map-matching algorithm developed in-house to take unintelligent GPS points and convert them into street centerline segments with all associated vehicle operation data. This algorithm runs around the clock, processing streaming GPS data in near-to-real-time, and serves up all output data to support live services used in internal and public applications. We also maintain a historical record of all raw and processed data for reporting and management. PVL runs entirely using Python and is supported by ArcGIS Server and GeoEvent Server running in Amazon Web Services. All our code is public on GitHub - all you need are GPS data from vehicles and a street centerline dataset to get started using it.
The Office of Sustainability's Energy Office wanted to create an application for citizens to see how suitable their rooftops are for solar panel installations. I used LiDAR imagery to calculate the amount of solar irradiance across all of Philadelphia County using average paths of the sun for each month of the year while taking into account the shadows that are cast by buildings and land features. The final product was a map showing solar energy by kilowatt-hour that I clipped and summarized by building footprints to show an estimate of how much solar energy could be collected when panels are installed in the most optimal positions and locations on each rooftop. The application can be seen here.
The City of Philadelphia releases a variety of open datasets to the public, including a number that are spatially enabled. OpenMaps provides the public with a simple application to explore a selection of open datasets from the City of Philadelphia. All layers provide easy access to the metadata with links to endpoints where you can download the data. This application was built using a webmap backend and using the in-house created Mapboard platform.
One of my first projects at the City of Philadelphia was the Stress Index. This storymap was created to display the results of an analysis including different potential stressors at a residential parcel level to help people understand the complex pressures on different neighborhoods in Philadelphia. I collected data from the American Community Survey, Philadelphia Police Department, the Center for Disease Control, and school districts to ultimately create a combined stress index. Using this applications you can explore datasets like asthma, distance to parks, drug crimes, heat stress, education levels, and poverty across Philadelphia. This tool has been used for the Community Schools Initiative to identify priority schools across Philadelphia to act as neighborhood resource centers.
In graduate school, I focused on using distribution modeling algorithms to predict the occurrence of wildlife and wildlife infectious diseases across large geographical ranges under both current climate and future climate scenarios (see my publications here). I wrote a suite of R scripts to use in conjunction with the biomod2 distribution modeling package to simplify it's use and generate a report with additional summary statistics and simple maps using Rmarkdown. More info here.