First came Hacker News, in 2007. It got developers around the world engaging in conversations about the hottest content on the web, day after day.
Yesterday a graduate student at North Carolina State University, Rohit Sivaprasad, started an online community for the subject he’s interested in: data science.
It’s called DataTau.
Sivaprasad doesn’t want DataTau to turn into a virtual land of obscure technical questions and answers, he told VentureBeat. Such threads might be better suited for sites like Stack Overflow. Instead, he’d like lots of people from all over the world to come to one site and talk with each other on a higher level.
DataTau users have already advised someone on how to nab a data science job, shared perspectives on using the emergent Julia programming language, and weighed in on the ongoing language debate pitting Python against R, among other discussions.
Such banter provides Sivaprasad with something he has lacked as a student. As a grad student in math for the past two years, he said, he has become interested in data mining and computer science. “One thing I noticed was that in mathematics, the whole feeling of community isn’t there,” he said. “One of the biggest things I’ve found when I was working on my research in pure mathematics was that mostly the only person I spoke to with my research was my advisor, and that once in a week.”
Now Sivaprasad and others who want to talk as they learn about and practice data science have a virtual venue in which to do it.
“I want people — if they work on something cool in data science, I want them to post it there,” Sivaprasad said.
Sivaprasad himself has built up much of his data science knowledge on his own, through online courses. He participates in Kaggle competitions and has made it into the top 500 contestants, he said. He has contributed to the scikit-learn machine learning toolkit for Python. And on the side, he’s been working on a startup called Predikt, to score people for their skills based on data from LinkedIn and other sites.
While the idea of a consistent, language-neutral, competitor-neutral water cooler for data scientists does sound appealing, the hard part will be keeping it going. Some Hacker News-like sites for other industries have shut down — one for quants lasted for less than three years. But sometimes it feels like everyone wants to be a data scientist these days, so maybe this one will have more staying power.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.