Skip to content
View andrewjtdunn's full-sized avatar

Highlights

  • Pro

Block or report andrewjtdunn

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
andrewjtdunn/README.md

Hey, I'm Andrew 👋

I'm a graduate of UChicago's Computational Analysis and Public Policy program. I have a passion for data science and experience in the consulting, financial services, and public sectors. My most recent work is as an AI strategy consultant and data science contractor.

I previously worked as a Data Science Research Assistant at UChicago's Crime Lab, where I supported the evaluation of randomized controlled trials by using natural language processing on video data. Before that, I led the Data Team at Financial Health Network as a Senior Manager.

You can see a few of my projects at the links below:

  • Civic Lens website. This open source site makes federal public commenting opportunities more accessible for the American public. I am lead backend engineer of this ongoing project, working with Django, PostgreSQL, and Python to manage the data pipeline for the site. Our code lives here.

  • Predicting Observed Poverty Levels in Costa Rica. Alongside three teammates, I tested different machine learning models, as well as techniques like ensemble learning, oversampling, and cross-validation to classify household-level poverty using observed characteristics. Our final model approached the top performing Kaggle models with a final Macro F1 score of .42. We summarized our approach and results in a summary report.

  • Comparing a Hardship Index with Evictions in Chicago. Using publicly available data, three teammates and I constructed an dashboard using Python and plotly to compare a constructed hardship index with instances of eviction. The resulting interactive data visualization allows the user to examine the index, its data inputs, and eviction data at the zip code-level in Chicago. We summarized our work in an explanatory report.

  • A/B testing of messaging in a fintech app. I designed and implemented A/B testing of behavioral design-informed messaging in a fintech app, resulting in 31% higher enrollment in automated savings program and 42% greater savings balances for users.

Other (private) projects include: building an app for a national nonprofit using an LLM to automate previously manual text extraction from hundreds of documents, constructing a data processing pipeline and testing different machine learning techniques to improve the accuracy of predicted policing outcomes by 40%, and improving a customer classification model for a B2B fintech company.

Feel free to connect with me on LinkedIn!

Pinned Loading

  1. uchicago-capp-30320/CivicLens uchicago-capp-30320/CivicLens Public

    Putting the public back in public commenting

    CSS 2 1

  2. Costa-Rican-Household-Poverty-Level-Prediction Costa-Rican-Household-Poverty-Level-Prediction Public

    A machine learning project to classify household-level poverty

    Jupyter Notebook 1

  3. uchicago-mscapp-projects/apwhy uchicago-mscapp-projects/apwhy Public

    30122-project-apwhy created by GitHub Classroom

    Python 3

  4. math-ml-project math-ml-project Public

    A project to build a machine learning classifier solely using numpy

    Python

  5. travel-maps travel-maps Public

    Maps made from phone location data of a month-long road trip

    Jupyter Notebook

  6. chicago-policy-review/Electric-vehicle-tax-credits-Nov-2022 chicago-policy-review/Electric-vehicle-tax-credits-Nov-2022 Public

    Data viz for electric vehicle tax credit article, Nov 2022