datascience@berkeley Summer 2015 Capstone Presentations
On Wednesday, August 26, datascience@berkeley graduates from the inaugural class presented their capstone projects — the culminating project of a two-year program — publicly in an online webinar. For their capstone project, the graduates were tasked with demonstrating the application of core data science skills related to problem solving, communication, influence, and management to a topic of their choice.
Many graduates utilized this project as an opportunity to solve problems that the average person faces everyday. From using predictive analytic strategies for a better user experience for San Francisco’s public transit commuters to manipulating large data sets of trending music patterns to determine the probability of individual songs being a “hit” in the rap music genre, among others, the MIDS graduates were able to quantify and solve real-world questions and obstacles. Project topics include:
- Decision-making tools for marketing electric cars to the individuals and regions most likely to buy them.
- Kelsey Clubb, Nikhil Kurup, Sean Y. Wang
- Using machine learning the team is able to predict a pitcher’s next pitch in baseball.
- Zachary Beaver, Jason Goodman, Joshua Lu, Alan Si
- Utilizing over 40 years of recording chart data, the team is able to predict what makes for a hit on the Billboard charts, as well as provide performance of similar entries.
- Tony Abraham, Nikhita Koul, Joseph Morales
Timely Trains – Better prediction for BART train departures
- Predictive analytical strategies allow this team to provide real-time, accurate understandings of the arrivals and delays of the BART train system in San Francisco.
- Milad Davaloo, Nitin Kohli, Elizabeth Peters
uConserve Home Energy Dashboard
- The uConserve dashboard is the product of analytics, visualizations, and targeted, actionable conservation recommendations to drive awareness and conservation of natural resources throughout the world.
- Lisa Kirch, Rahul Bansal, Nital Patwa
- Using comparative analytical strategies based on one’s current zip code, the team is able to aid in the home buying decision providing similar target areas in other metropolitan areas.
- Sharon Lin, Judd Heckman, Thiyagarajan Shanmuga
- Wilson provides a transparent look into strategizing Lending Club investments coupled with risk management.
- Nathan Rogers, Erin Boehmer, Kunal Shah, Tao Mao
Watch the Capstone Presentations
“We created Wilson, [an online lending club database] which allows you to see exactly which loans we are predicting will generate higher yield for your portfolio, [and] allows you to keep control over what you’re investing in, keeping it very transparent.” – Erin Boehmer, MIDS Graduate
Citation for this content: datascience@berkeley, the online Master of Information and Data Science from UC Berkeley