The Land of Oz Ozzie Liu

Liv Buli of Next Big Sound Talks to Metis Data Science

Every student at the bootcamp were staying late and working through the night on their movie data project, so it was a very nice break when our guest speaker Liv Buli came to speak to us on Thursday night…

Liv is currently a data journalist at Next Big Sound, and she shared with us on “Story Telling + Data Science + Data Visualization = Data Journalism”. She shared with passion and a no-frills attitude about her work as a data scientist and gave some great advice on the importance of good data visualization to tell our story.

A chart like this is bad. Percentages on a pie chart that adds up to more than 100? fox pie Fox News

But a chart likes this tells a story and much more: napoleon russia Wikipedia

She provided a great example of a project that was 2 years in the making since its inception. It’s an incredible visualization that explains Next Big Sound’s work at benchmarking artists’ social media engagement with other artists within their popularity “bucket”. You should be able to find it on Next Big Sound’s website soon.

Her advice for upcoming data scientists? It sucks sometimes.

Not just the projects are difficult and long, but there are times when our investment of time and resources really do not result in presentable or interesting insights.

At the end, she offered 4 points for all data scientists to keep in mind:

  1. Don’t be afraid to murder your darling. The idea here is to proceed with objectively and without sentiment. Sometimes we can get attached to our hard work and prevent us from changing focus to tell a better story or even scraping it entirely.
  2. Data is not going to be perfect. Data out there is always going to be messy, incomplete, even discombobulating. That’s just reality. So there’s no point in waiting for the perfect dataset, there are still insights to gather. But at the same time, great data may also not lead anywhere, which leads to the fact that -
  3. Sometimes there just isn’t a story there. So maybe there’s a different question to be asked. Or a different approach. And finally:
  4. Can you explain it to a four year old?