Our company is interested in researching the different options for synthetic data generation, with Python libraries as a starting point. Applications of this work span the variety of services we provide in data creation, management, and analysis to potential clients in the social impact sector.

We would like to collaborate with students to research how best to compare the available Python packages for synthetic data generation, which typically involves use of statistical methods alongside generative models (i.e. generative adversarial networks, variational autoencoders) and sometimes conditional generative models. Students will create a summative white paper style document as well as use that as a basis for a Jupyter notebook that rigorously and holistically demonstrates the findings.

This will involve several different steps for the students, including:

• Conducting background research on how existing Python synthetic data libraries were created and set up.
• Analyzing our current services offerings in data creation, management, and analysis.
• Researching how existing Python synthetic data libraries could be applied for potential clients.
• Developing a summative white paper that provides unique outcomes or insights into our application for potential clients.
• Providing a Jupyter notebook that can be used to guide our exploration of next steps for potential clients.

By the end of the project, students should demonstrate:

• Understanding of the available Python synthetic data libraries
• Understanding of the latest AI / ML techniques related to synthetic data generation
• Identification of ways in which synthetic data AI / ML techniques can be applied to our company

Bonus steps would include:

• Providing comparison to synthetic data libraries in other programming languages

Final deliverables should include:

• A final white paper on the synthetic data libraries' setups, the problem(s) each library solved compared to the others' methodologies and approaches, comparison relative to each service our business provides, and recommended next steps regarding potential client work.
• Source materials such as code and the Jupyter workbook(s) guiding our exploration of next steps for potential clients.