The digital home of the Michigan Data Science Team. Inquire below.
Competitive Data Science at Michigan
Meeting Times for Winter 2018
- Project Team Meetings: Thursdays 5:30pm to 6:30 in BBB 3725
Check the calendar to confirm there is a meeting!
Social events occur every two weeks, typically after Thursday evening project meetings. The last Thursday of every month an MDST group attends Arbor Brewing Company trivia. Other social events vary from game nights to the annual barbecue to tailgates; you can find up-to-date information about social events on the calendar and in the weekly emails.
Prospective Members should sign up here to join MDST.
We do not offer a tutorial series during the winter semester. However, last semester’s tutorials series can be found in entirety online. The videos are available at http://leccap.engin.umich.edu/leccap/site/qrbkcawjkoyut2llnkr and the code is all available on github as jupyter notebooks at https://github.com/MichiganDataScienceTeam/tutorials.
ImageNet Replication Project
We will replicate systems from recent deep learning papers and apply them to the ImageNet image classification dataset. You and your teammates will first pick one paper to replicate, and then try to recreate the results from the paper. After you recreate the original results, you will run at least one new experiment using your replicated system. This new experiment may aim to further analyze the system you’ve replicated, or it may attempt to improve upon the previous system by changing or adding a new component. The goal of this project is two-fold: (1) to familiarize yourself with current deep learning frameworks and literature, and (2) to practice running focused experiments on large datasets. This is a purely educational project to help you learn new skills. This project launched in early February, and new members are welcome to join pending project leadership interest.
Turning The Corner
MDST is partnering with Data Driven Detroit to continue their work with the Turning The Corner project, which aims to be able to predict neighborhood change and displacement in Detroit. If you’re interested in joining this project or would like to learn more about it, check out the slack channel (#turning_corner_gen) for more information. This project launched in January, and new members can still be accommodated.
NFL Free Agent Prediction Challenge
This event kicked off in November, registration closed in December, and the live model evaluations will take place in April during the free agency period.
In this competition, student teams at the University of Michigan will use historical free agent data to predict the value of new contracts signed in the 2018 free agency period. These predictions will be evaluated against the actual contracts as they are signed. This competition is organized by the Michigan Data Science Team (MDST), in collaboration with the Baltimore Ravens and the Michigan Sports Analytics Society (MSAS).
The Michigan Data Science Team (MDST) is a competitive collegiate data science team at the University of Michigan, Ann Arbor. Together, we compete against professional and amateur data scientists from around the world in online prediction challenges.
Competitive data science has become increasingly prominent in the past decade with the immense popularity of high-profile competitions such as The Netflix Prize. Now, online venues such as Kaggle, DrivenData, and Quantopian, among others, provide platforms for data scientists around the world to make impactful contributions to a huge variety of prediction problems while competing for cash prizes. Previous competitions have explored prediction problems in healthcare, particle physics, finance, and countless other domains, and have involved many types of structured and unstructured data.
We are an organization at the University of Michigan in MIDAS that is looking for dedicated students who are interested in taking part in data science competitions. Ideally, students will have a strong background in computer science, mathematics, and statistics, but all students are welcome to participate. We will be meeting on a weekly basis throughout the year, where students can share strategies and give tutorials. Anyone is welcome to attend these meetings! For your hard work and dedication, we will be offering internal prizes to the students who achieve the highest performance on our prediction challenges.