Kristin Chang

Kristin Chang

About Me

Hi! I'm an experienced GIS and spatial data analyst with professional training in city and regional planning. I am passionate about leveraging the potential of geospatial insights to facilitate data-driven decisions and create more sustainable and resilient urban environments.

I am based out of Oakland, California and am looking for opportunities to bridge my technical expertise with my interest in transportation and climate action planning.

Check out some of my projects below or contact me to learn more about my work!

Contact

Resume

My Projects


Spatial Data Analytics


Thumb

Historic road conditions data and the city of El Paso's Pavement Condition Index (PCI) are input into a Random Forest model to predict future PCI scores for each road segment under the city's jurisdiction. These predictions are used in the project's counterpart, a scenario planning tool, created during the MUSA Practicum course at UPenn in Spring 2022 for the city of El Paso's Planning Department.



Thumb

This project utilizes a generalized linear model to predict areas of future flooding in Denver, Colorado trained on past data from the City of Calgary in Canada. The predictions are intended to hypothetically inform Denver City Agencies such as Public Works, Public Health, City Planning, and Emergency Management to better prepare for future flooding disasters.



Thumb

My first deep learning models! This project uses Convolutional Neural Networks (CNN) to classify EuroSAT images by land use and land cover. I explore the results of different models using both RGB and MultiSpectral images.



Thumb

This toolkit for a Green New Deal provides digestible summaries of complex datasets for policymakers advocating for environmental justice in regions most impacted by fossil fuel production and climate change. The project consists of a Sentiment Analysis of tweets using the hashtag #GreenNewDeal, a Spatial Analysis of Fossil Fuel Infrastructure and Air Pollution, and an analysis of Air Pollution levels in Louisiana from 2001 to 2014 - all with interactive charts!



Thumb

This project utilizes drone imagery of Rwanda to hypothetically inform the country's crop management decisions. Machine learning and deep learning techniques are implemented to classify each image as one of six crop types to quickly and efficiently obtain a snapshot of Rwanda's food supply at a given time.




Thumb

A custom data dashboard hosted on heroku summarizing reported outreach activities in Orange County for the 2020 Census. This tool was created for Charitable Ventures through the Good Work Collaborative to better visualize the Orange County census network and to serve as a catalyst for future organizing.






City Planning


Thumb

The St. Thomas Resilience Plan was created in the 2nd year City Planning Studio for St. Thomas located in the U.S. Virgin Islands to address the island’s vulnerability to climate change and sea level rise. Notable contributions include construction of a St. Thomas specific Social Vulnerability Index (pg. 24-25), Demographic and Socioeconomic analyses (pg.26-27), mapping traffic patterns (pg. 68), and graphic design and layout of the Existing Conditions section (pg. 24-29).



Thumb

Predicted pavement conditions power this scenario planning tool that allows the city of El Paso to better allocate its capital improvement funds for road repair projects. City Planners are able to set a threshold for pavement condition scores and view these road segments in conjunction with other areas of interest (e.g. Equity, Traffic Congestion, Road Safety, or Flood Risk). This tool was created for the city as one of two parts of the MUSA Practicum course at UPenn in Spring 2022.



Thumb

FreshFridge is an Arduino powered kit prototype created to better inform community fridge managers. Various sensors collect data on the status of the fridge and a WiFi shield enables realtime database collection via Google Firebase. Notable contributions include graphic design, wiring parts for shelf capacity, and writing Arduino (C/C++) code for shelf capacity and recording multiple concurrent timed events on the same microcontroller.