Data Science Practicum
The Data Science Practicum (DATA-793) is the capstone experience for the MS in Data Science and provides assistance to faculty, government organizations, and companies. See examples completed Student Projects.
Students entering the practicum have completed coursework in statistics, regression, and R for data science, along with completed or ongoing coursework in statistical machine learning. Practicum students are ready to put their visualization, analytics, and data modelling skills to work on live projects.
Practicum Coordinator:
Maria Barouti, barouti@american.edu.
Call for Projects: Faculty & Staff
Let our advanced students help with your data:
- Use the Faculty Project form to provide your project title, description, AU email, and required skills — please use our course listings to help specificy required student skills or classes.
- Under the guidance of our faculty, students in the Data Science Practicum (DATA-793) or other advanced research courses review available projects for best fits and contact you.
- You and the student(s) agree on a plan for work on the project.
| Organization | Sponsor | Title | Description |
|---|---|---|---|
| AU CAS Math-Stat | Richard Ressler: rressler@american.edu | AU CAS Undergraduate Advising Chatbot | Expand and refine a prototype Retrieval-Augmented Generation (RAG)-based chatbot to support undergraduate advising in mathematics, statistics, and data science programs. This involves updating and developing new python code to scrape publicly accessible webpages from American University to collect information on academic programs, policies, courses, and study abroad opportunities and refining the logic and code for embedding and retrieval to improve the ability of the generative AI models from HuggingFace to produce relevant answers. |
| AU CAS Math-Stat | Nimai Mehta: mehta@american.edu | Land Fragmentation, Use, and Consolidation in India | Access to land remains a strong predictor of equitable growth and livelihood for millions of households in South Asia, and an important basis for gender and inter-caste equity. Land holdings across India have displayed alarming levels of fragmentation over time. As a result, access to productive plots of land remain out of reach for most households in both urban and rural districts of India, and for public purposes. At the same time, the continuing lack of clear and secure titles to land along with a conflicting set of land-use and land-transfer laws at the local/state level have worked to aggravate problems of land scarcity and access. The response by the state has been two-fold. At the local level, various states have passed Anti-Fragmentation and/or Land Consolidation laws that impose further restrictions on land transfers while providing various incentives/schemes to encourage consolidation. At the federal level, the state has dramatically widened its powers of eminent domain to consolidate and secure land needed for large private and public projects. Stuents will use a mix of data science and LLM (Large Language Model) tools to extract and analyze the textual – legal, institutional, and financial - data associated with land use records. |
| AU CAS Math-Stat | Zois Boukouvalas: boukouva@american.edu | Interpretable Bias Detection in Deep Neural Networks via multi-modal fusion techniques | Context and Goals: This project aims to develop principled and interpretable methods for detecting and mitigating bias in deep neural networks by leveraging advanced multi-modal fusion techniques. By jointly analyzing neuron activations across layers, the framework will identify latent components and representation subspaces that persist through the network and may encode bias. This approach will enable the localization of such subspaces, facilitate interpretation of their alignment with diverse dataset characteristics, and guide the design of targeted interventions to improve model fairness, robustness, and transparency. The aim of this project is to bridge unsupervised representation learning, statistical signal processing, and responsible AI principles to produce tools that generalize across architectures and data modalities. Student Support: The project can support two students, working collaboratively on the theoretical design, algorithmic development, and experimental evaluation of fusion-based bias detection and mitigation methods. Students should have strong Python programming skills, practical experience with machine learning, and a solid foundation in statistics and data analysis. |
| AU CAS Math-Stat | Mary Gray: mgray@american.edu | Analysis of Discrimination Lawsuits Regarding Faculty Salaries in the US. | A research team is analyzing studies on faculty salaries, and this project is focused analyzing how salaries are analyzed if they go to court, in particular for remediation issues. The student would search for court cases of salary discrimination in the last 5 to 10 years that have been settled in favor of the plaintiffs with emphasis on university faculty and staff. These would be Title VII, Title IX, age discrimination cases for the most part. Then they would have to collect the data and analyze it in some detail to assess the statistical analysis used in supporting the case, (but not the legal issues). The student would then compare the methods used to arrive at the decisions of the courts. In particular, of course, the issue of whether AI has been used or at least mentioned in the case. Deliverables would include a data set and general listing and review of cases with an analysis of differences and similarities in the outcomes and associated factors. |
| AU SPA Government | David Lublin: dlublin@american.edu | Inclusion by Design | Minority inclusion is at the center of not only creating more equal societies but also democratic stability and preventing ethnic conflict. Yet scholars have found no solution to the knotty problem of measuring inclusion across countries. This problem limits our ability to learn how to design political institutions, such as the electoral system or federalism, to enhance minority inclusion more effectively. My solution to this problem centers on estimating minority electoral support for governing parties in legislatures (and for winning presidential candidates where the president serves more than a symbolic role). Using this information, one can also estimate the minority share of the government’s (and the president’s) electoral supporters. Towards that end, I am taking a multipronged approach to estimating voting behavior by different groups, relying on both ecological inference and polling data. Prerequisites I need students who are very comfortable (1) locating polling data, and (2) getting key descriptive stats properly weighted out it. Appropriate skills in statistical packages are helpful. I have a grant to pay students over the summer who are interested. |
| Strata9 | John Barnshaw: johnbarnshaw@strata9.com | College Promise Program Analysis | This project builds on previous work to analyze and integrate Promise programs into the Strata9 database, and from there, analyze the economic data they have collected over the summer to determine areas of opportunity to help Promise programs better understand the labor market. It is estimated that 40% of this project would be joining and data collection, and the remainder would be analysis and reporting. |
See previous projects.
Top image credit: Fabio, Unsplash.