Gregory Mitchell Jr
Current student at the University of Chicago studying Computer Science and Public Policy
Current student at the University of Chicago studying Computer Science and Public Policy
GPA: 3.85
Select coursework: Software Engineering, Machine Learning, Linear Algebra, Mathematical Foundations for Machine Learning, Statistics, Discrete Math, Algorithms and Data Structures, Cloud Computing, Computer Systems
GPA: 3.81
Select coursework: Econometrics I & II, Microeconometrics, Mathematical Economics, Spatial Demography
GPA: 3.98
Select coursework: Microeconomics, Macroeconomics, Urban Economics, Development Economics, Environmental Economics, Gender and International Development, Hydrology, Climate Science, Research Methods
Served as a Teaching Assistant for Computer Science with Applications I (CAPP: 30121) and Computer Science with Applications II (CAPP: 30122). I led discussion sections where we reviewed computer science fundamentals and computational thinking. Held office hours to provide guidance on weekly programming assignments. Provided timely and thoughtful feedback on assignments.
Python · SQL (SQLite, PostgreSQL, SQLAlchemy) · FastAPI · pytest · Docker · Google Cloud (GCP) · Jira
Assisted in the development and deployment of Ask-a-Metric, an AI-powered, text-to-SQL tool designed to provide real-time data analytics to non-profit and government leaders via a simple WhatsApp chat interface. In my time with IDinsight, I built core features to improve language support, database support, user feedback collection, and overall performance. I also developed automated tests and LLM output validation pipelines to prepare for SaaS deployment.
Stata · R/RStudio · ArcGIS · tableau
Provided research and analytical expertise to 10+ nonprofit organizations and governmental agencies to promote social justice. I was responsible for managing ongoing client relationships and led the research and writing of 3 major client deliverables. These deliverables included a literature review for the United Way, a housing needs assessment for a county government, and multiple Tableau dashboards. I also served as a quantitative and technical resource to my organization, consulting and providing guidance to multiple projects including developing a quantitative program evaluation plan for a youth leadership initiative in New Orleans. Additionally, I was responsible for leading causal inference and econometric analyses in Stata and R to support housing and community development efforts.
Stata · R/RStudio · tableau · Maptitude
Supported economic experts in providing quantitative and qualitative analysis on antitrust issues across multiple sectors, but primarily healthcare. In my consultant role, I processed and analyzed millions of observations in Stata across dozens of public and proprietary data sources. In the Healthcare practice at Bates White, I constructed spatial analyses to evaluate market concentration in the healthcare industry for regions across the United States, using CMS and insurer claims data. I developed scripts to automate the data analysis and visualization generation to deliver faster and accurate results. Ultimately, I led the creation of 100s of maps, charts, and other visualizations in R, Tableau, and Excel to communicate to diverse audiences.
Stata · R/RStudio
Analyzed the Federal Reserve’s Survey of Household Economics and Decisionmaking (SHED) to identify disparities among young workers in the areas of credit and banking access to prepare for an internal report.
Worked with the United Way and a small team to assess the transportation needs of residents in Centre County, PA. Developed and executed a research plan to interview and survey over 200 human service organizations and businesses in the region.
Python · Pandas · NumPy · scikit-learn · NLTK · FastAPI · JavaScript · AWS (S3) · Heroku
Used Support Vector Machines (SVM) and Truncated SVD for dimensionality reduction to build a lightweight model to predict cyberbullying and hate speech with between 80-95% accuracy and precision, depending on the category of cyberbullying. Processed tweets and applied sentence embedding techniques (word2vec and TF-IDF) for feature extraction. Deployed the model as an online interactive demo using FastAPI, JavaScript, Heroku, and Amazon Web Services (AWS) S3.
Note: The demo is being run at the demo link above, however, due to the limitations of the student-tier of Heroku with a maximum of 512 MB of memory, there is some instability. Please reach out if you run into issues using the demo.
Go (Golang) · Bootstrap · HTML · CSS · JavaScript · Heroku
Developed this website using a Go server with a Bootstrap theme and custom CSS. This website is hosted on Heroku.
Go (Golang) · JavaScript · HTML · CSS · Heroku · PostgreSQL
Developed a web application to help users find relevant businesses equidistant from where they and their friends live. I built a RESTful server in Go (Golang) and a dynamic front-end in JavaScript with an interactive map and business search results. I also developed a Go module to generate hexagon tessellations using computational geometry for API querying.
The idea for this project came from my time living in DC, where I often found myself trying to coordinate with friends in Virginia, Maryland, and other parts of DC. This application allows users to input the addresses of their friends and search for restaurants, bars, etc. that are in between them and their friends.
This project was my first foray into Go and I built the initial project in about 2 weeks after I returned from Kenya following my internship at IDinsight.
Demo is live! I am in the process of fine tunning and making optimizations to the app (including making it more mobile friendly), but v1 is now available.
Python · NumPy · Pandas · TensorFlow · OpenCV · Matplotlib
Built a convolutional neural network to classify restaurants into price categories using Google Street View and Yelp data. I collected and processed image data from the Yelp and Google APIs to prepare them for the machine learning model.
Github code forthcoming.
Python · SQL (SQLite) · Ollama · Mistral 7B
Led the development of an AI chatbot (Mistral 7B) designed to generate summary statistics and interactive maps from SQL data. I helped manage the overall product, develop the strategic plan, and coordinate the team, tasks, deadlines and project requirements. I also served as a coding resource for fellow team members.
This project was my first time working with LLMs and using their ability to help convert natural language queries into SQL code. My work on this project fed directly into my work at my summer internship in Kenya at IDinsight.
Github code forthcoming.
Stata · R/RStudio · ArcGIS
Environmental, green, or ecological gentrification is the process whereby improvements in local environmental amenities—such as new green spaces or the cleanup of a locally undesirable land use (LULU)—either cause or exacerbate gentrifying processes. This a direct threat to the progress of the environmental justice movement whereby residents of low-income neighborhoods and communities of color face the highest possibility of displacement, negating any benefits of environmental improvements for these populations. I developed spatial econometric models using Stata, R, and ArcGIS to uncover causal evidence of the displacement of Black households following pollution cleanup in Allegheny County, PA from 1990-2010.
This is my BS/MS thesis from the Pennsylvania State University. My work won the 2020 Outstanding Masters Thesis Award from the Northeastern Agricultural and Resource Economics Association, and I ultimately presented on this work at the 2023 Urban Affairs Association Conference.
I absolutely love music. I primarily grew up listening to R&B and Hip Hop, but I've consumed a lot of music over the years. When I'm not at school or doing work, there's a high probability that I'm at a concert, club, or just jamming out in my apartment. Below are some playlists I've recently created and been listening to a lot 😁