Datahack is a one-day Spring hackathon co-hosted by Texas Convergent. Participate to win prize money and network with industry leaders.
Past topics are listed below:
Now that the NOAA has been defunded, an insurance company, InsuraCorp, is seeking consultation on the use of machine learning models to predict the outputs of larger climate models. They have provided a dataset from a custom climate-damage model to test teams on their ability to forecast both weather and the resulting damage. Teams are given a dataset containing pressure, temperature, wind speed and direction, along with civilian damage over time.
Your task is to consult the Houston Astros on selecting free agents from a pool of 150 players, using 10 seasons of data (1620 games per team). You need to deliver two items: forecast next season's hits for each free agent, and develop a bid strategy allocating $200 among players, with the understanding that you'll acquire the top 3 players you successfully bid on. The Astros' primary goal is to maximize wins next season with their new free agents, but you should also consider alternative goals in your consultation.
Your DataHack 2023 team at a private equity firm is tasked with evaluating an e-commerce platform investment, focusing on the issue of fake reviews. Using a dataset of genuine Amazon reviews, fraudulent reviews, and AI-generated reviews, your team must address three key questions for the firm's executives: the effectiveness of fake review detection, the importance of user metadata and textual content in this process, and the potential for large language models to create more sophisticated spam reviews in the future. This analysis will help inform the firm's investment decision and strategy regarding the e-commerce platform.
Your task is to advise a private equity firm on whether to invest in Upworthy, a media site known for viral, clickbait-style headlines. Using data from A/B tests (headlines + images), user engagement over time, and country-level usage, your team must assess Upworthy’s claim that it abandoned clickbait. Deliver insights on how the company has evolved, and evaluate whether those changes signal sustainable business growth. You may apply NLP to measure clickbait, use time series analysis to detect shifts in behavior, and explore ethical, financial, and user implications.