Texas Datahack
🏅 Competition Overview
Texas Datahack was co-hosted by MLDS (Machine Learning & Data Science) and Texas Convergent, two organizations at UT Austin, and was sponsored by Oracle and Chevron. It took place in October 2022. Participation was open to all students, and 22 teams competed. We had ~6.5 hours from receiving the challenge information to presentation submission. My team received first place.
The Challenge
Your DataHack team is part of a data science group at a private equity firm. Leadership is interested in making a significant investment in Upworthy, a viral news publishing & aggregation website. However, in the past, Upworthy has been disparaged for relying on clickbait to generate clicks, exposure, and signups. In response to this, Upworthy pledged to stop creating clickbait headlines and articles. Your firm’s executives want to invest in Upworthy, but they are skeptical about whether Upworthy genuinely changed their publishing practices or not. Using a rich dataset of A/B tests from Upworthy, your team seeks to answer the following questions:
- Did Upworthy really change their clickbait publishing practices?
- What was the impact of that change?
- Do you recommend an investment in Upworthy?
Challenge description is taken from the Datahack Challenge Packet. Further details can be found in the two files below.
🎬 Presentation
https://www.canva.com/design/DAE9aHDfGBk/8T6Zf5r14UPSl273W-EZsg/view
https://www.canva.com/design/DAE9aHDfGBk/view
⌨️ Code
Please note that the code is very messy due to the time constraint of the competition and may not be particularly readable.