Throughout the summer, Skandalaris Center Summer Internship Program participants will be writing guest blog posts about their internship experience. Following is one such post. 

Amazing Experience Involved with Analytics and Teammate

By Yujia Wang (GB – Customer Analytics Program)

I am working for the Advocado as data scientist during the Summer Internship Program. I really appreciate this opportunity. It is a great chance to feel the working environment and pace. Everyone is able to come up with their own ideas and there is always laughter and talks in the office. Everyone at Advocado is really nice and is eager to offer help to me. On the first work day, they had a welcome lunch for me with other stuff, which gave me opportunity to know people from other departments. They not only help me at work, but also give me advice on life including the choice of car and lunch. For example, when I didn’t know the data structure for the database, a member of the engineering team provided me with the solution and answered my question patiently, which really impressed me and encouraged me to chase the working pace. People in this venture are passionate and kind. It looks like they are big family or have been old friends for a long time, although they actually haven’t known each other for a very long time. The company has a daily stand-up huddle and engineering meeting routine, which pushes everyone to complete their work and communicate with one another every day.

As for the work, my job is related to data analysis for customers and advertisement, which really interests me. It’s involved with observing people in normal life and knowing about individual’s habits and other attribution. I am very glad that I am able to apply all I’ve learnt about customer analytics in Olin to the real practice. For instance, I analyze the data on Google AdWords by using digital marketing knowledge. One of my favorite projects is that I fit the deep neural network model by Python to advertisements data to predict advertisement online performance and find KPI (key performance indicator) and score for market cluster by random forest model. It’s interesting but challenging. It’s amazing since it’s real data and you can understand people better through these figures. But it took me lots of time to testify the hypothesis and modify the model. I can never know whether it’s the best solution since nobody has the correct answer in reality. I can only count on myself. It is a little frustrating for me but then my boss praised my work, which really inspires me and makes it worth it.