This is a project that I completed individually as a part of the Forge program (https://joinforge.co/launch-ds) that I eventually taught. It was completed over the course of one week. You can find my Jupyter notebook and two Tableau workbooks in the Git Repository as well as the slides I presented to my Forge data science cohort, instructors, and Forge alumni in the workforce.
The Project Brief was as follows:
Source: US DoE and others (via Tidy Tuesday, https://github.com/rfordatascience/tidytuesday/blob/main/data/2020/2020-03-10/readme.md)
- Data dictionary is available in above link.
Employer and task:
- You’re a data scientist at collEDGE, a startup that helps high school students from all backgrounds focus their application efforts towards schools that are the best possible fit for them. collEDGE is interested in you leveraging the included data to provide data-driven college recommendations to students, especially for those they might not have heard of before.
- Note: as this task is a bit broader than the rest by nature, feel free to create one or several “case studies” in addition to (or perhaps instead of) your general recommendations, showing what schools you would recommend to collEDGE is also interested in any general insights about trends in campus diversity, expected salary, tuition, etc. that you’re able to find, as they market to students primarily through blog posts on these topics. Guiding questions: These are meant to help point you in a productive direction, but remember that your ultimate task is stated above.
- What characteristics are desirable in a school? How do these vary across different types of students, and how can you account for this diversity of backgrounds and goals in your recommendations or case studies?
- How do your rankings or recommendations compare to the more popular ones published by Business Insider and the like? Why should students prefer collEDGE’s rankings?
- What’s going on in higher ed right now?
Deliverables:
- A polished, 5-7 minute slide deck presentation featuring high-quality, well-motivated, well-presented visual and analytic findings.
- An interactive chart or dashboard (in Tableau, Plotly.py, or other similar tool) allowing users to conduct their own investigation into your research question.
You can find my code, slides and Tableau Workbooks in this GitHub repository. Note that Tableau Workbooks are packaged with extracts of the data, meaning if the data source is updated, they may not reflect updated source of data. This is because Tableau Public for Students (free) does not support live connections as of June 13, 2025.