If you haven’t check out Datafest 2016 Reflection go here now: http://ugrid.info/2016/04/01/Datafest_2016_Reflection.html

I would like to take a moment to thank my awesome team members (Off-The-Grid) and the wonderful judges to winning a medal in the Overall Best in Show Award AGAIN at the 5 College DataFest! We were 1 out of 3 teams to receive this award out of the 30+ teams that participated. It was a blessing… to work with large data again within a short time limit.

What I've Learn From This Entire Experience:

  • Clear up calendar space for Datafest.

  • A well crafted story and having a silver tongue is an important to a data science competition.

  • As data scientist, our job is to make credible and useful recommendations.

  • Knowing your audience, again stop and think about who are you presenting your findings to.

  • When working with big data it’s important to have data wrangling skills (one of the most important tool in your arsenal).

  • Tools such as Amazon Web Services can be a great help to boosting CPU or memmory performance.

  • Software such as Tableau is great for making data visualization on the fly. It is also quite handy for data exploration.

  • Unique ideas are a double edge sword; there were groups with excellent presentations, but fell a bit short because they were lacking a strong data analysis.

In the end, when looking at Expedia’s data; they gave us clicks rather than people. A story would’ve been better driven if those clicks transformed into potential customers for the analysis. Another suggestion is to bring in use of outside data, there were alot of key features that were missing in the data presented. I believe that we won this time around was because we had strong summary statistics and a relatable business insight for our story which may have earned us additional brownie points. Key phrases such as “every customers only have a 17 click window to become conversions”, gives the targeted audience an insight of where to address a problem.