Data Analysis Mini-Hackathon Review

Didn’t make it to the first Mini-Hackathon with Galvanize and No problem.

This last Friday (Feb. 17th, 2017) Galvanize and teamed up to put on a mini data hackathon in Austin, Texas. Hosted at the downtown Galvanize campus and based on the “product hunt” data set, the hackathon participants — 55 data enthusiasts of all skill levels — teamed up into 11 groups of three to five to figure out what makes a pre-product launch on product hunt successful. The product hunt data set included two years worth of post (product pre-launches) and user data from the product hunt website. Groups worked tirelessly (excluding short breaks for pizza, of course) from 7:00 PM to 10:00 PM using data analysis techniques ranging from descriptive statistics and data visualizations to NLP feature-based regression and machine learning clustering techniques. With only three hours to get oriented to the data and deliver an awesome presentation, it was amazing how successful the teams were at identifying interesting questions and producing compelling analyses to support their answers. But of course, everyone wished they had more time!

Special Thanks

  1. Special thanks to all the team members from and Galvanize that made this event possible.
  2. An extra special thanks to all the participants who made the Mini-Hackathon a HUGE success!
  3. Finally, special recognition is due to our prize winning teams. It was very difficult to select from all the outstanding teams — all the presentations were excellent — but our winner selections are as follows.


The winner of the “Best Visualization” category goes to team 2: PJ Shetty, Alona Varshal and Ed Solis. They produced a “story” of sequential visualizations exploring the differences between highly up voted posts and posts that don’t generate a lot of interest on product hunt.

The winner of the “Best Answer” category goes to team 10 — aka Jake Schmidt, Maurya Avirneni, Fred Nugen and Jacky Gaschot — who sought to identify predictors of user activity level on product hunt and identified positive relationships between social media presence and product hunt posting, commenting, and up voting activity.

The winner of the “Best Question” category goes to team 5’s Mohit Motiani, Nish Patwa, and Ala Raddaoui. They came up with a panel of objectives for business insights and began work using unsupervised learning techniques to identify target demographics on product hunt, of most interest to marketers seeking to leverage the product hunt website.

A most honorable mention goes to Prakhar Jain, Trevor Thompson and Tomas Pihrt of team 1, who identified which times of days were optimal for launching products on product hunt. They used a stratified bar chart that examined the monthly growth of topic categories, and began exploring techniques to predict churn in an effort to better direct product hunts efforts in maintaining their user base.

Congratulations to all who participated and we look forward to seeing everyone again at our next Data Hackathon!

— Written by Galvanize Data Science instructor, Scott Schwartz on behalf of the Galvanize and teams.

Like what you read? Give Kelly Holt a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.