Data-Driven Storytelling
Here are some examples of the projects I've completed, and how they were presented. In a group project, I often take the lead in the timeline, delegation, and visual/written tasks, in addition to my share of the technical analysis. These project management responsibilities were a large component of my career prior to graduate school, and was thus an easy transition.
Predicting Repeat Listens of Songs: Presentation
This was a 20 minutes presentation, given as approximately 50% through a semester-long group project. As the first team to present, our concluding slides focused on our tentative upcoming steps that were modified based on professor and peer feedback. The final submission was a report, found below.
Predicting Repeat Listens of Songs: Report
This is the report from the aforementioned project. As somewhat of the "team-coordinator," I created an outline of the report and delegated sections based on what people knew the most about. I edited the final report, but asked other individuals to be a second editor for sections they did not write. Some of the sections I originally wrote include the Abstract, Forum Review, Lessons from Individual Models, Accounting for Time Series.
Text Analytics of Beer Reviews
This project was completed over 2 weeks by my team in a Text Analytics course. We scraped text and ratings for craft beer reviews on beeradvocates.com, then derived business insights for craft breweries. I added an interactive element by creating a Qualtrics survey for classmates on 10 commercial beer preferences (Heineken, Lone Star, etc.) and used cosine similarity to recommend craft beers that were closest to someone's survey responses.
​
This was a bit of a playful project, and our professor was so proud that he posted several of my group's slides to his Facebook page (after asking our permission), praising our work and sharing it with his network.
Hit News Science: UT Newspaper Network Analysis
In this project, my team scraped data from UT's student newspaper, the Daily Texan. This included the article's text, date published, author, auto-generated tags, and the website link. We had over 6,000 articles from 2009 to February 2018, and were able to extract topics from the text