POSTED: June 3, 2020
Much of the criticism regarding “toy data” or “classroom data” says that they are too clean, too perfect, and not representative of real-world data. There is much merit to that criticism. I remember the first time as a freshman in college when I tried opening 2GB of data in Excel … keyword “tried.” I had to do a hard shutdown on my laptop immediately and I refused to touch it for two hours because I was afraid of what I just put my computer through.
I am a recent graduate from Miami University and as I reflect on how far I have come with data since that moment, I attribute the immense growth to the countless learning opportunities I have had since then. Several GBs of messy data does not scare me anymore. One of the most important contributors to this growth was all of the exposure to real-world clients I had in college. I am grateful that my “classroom data” often was provided from such clients as the local county health commissioner, the local county coroner, experiments from other departments within the university, and even companies. These opportunities left me more confident with my data science abilities and excited to continue working with clients to solve problems using data.
Now that I am in the industry, I wanted to help give current students that same opportunity. We at Covail believe in the power of collaboration so we facilitated a client project for a Miami University Business Analytics practicum class. We partnered up with Pelotonia to provide data and a problem for the students. Pelotonia hosts an annual bike ride to raise money for cancer research in Columbus, Ohio. A couple of years prior, Covail did some work for Pelotonia relating to rider retention. We decided this would be a great project to revisit, so we tasked the students with building a model to predict rider retention and an interactive application for the client to view, download, and explore results. The application and model will help Pelotonia take proactive actions to retain as many riders as possible in order to raise more money for such a great cause. The students found variables related to previous years ridden and variables related to the donations a rider fundraises to be the most important in predicting rider retention. Note that no personally identifiable information was provided to the students.
The students were not only expected to deliver a reusable solution that will provide predictions for every previous rider for new years of data to come, but also ask questions and interact with the client as needed to solve the problem. I was so impressed with the professionalism of the students and proud of the deliverable they produced for the client. I look forward to these students starting roles as data professionals with some experience to excel them forward in their career. Colin Amy, a senior ready to start his career, states, “The opportunity to have my school projects be from real companies is an amazing experience.” He continues saying “When fake data is used in projects there always seems to be an easy, direct solution; the problems don’t present themselves the same way they do in the real world. Problems with real companies and data give me an opportunity to develop realistic problem-solving skills.”
Collaborate with us at Covail to see how we can help you leverage artificial intelligence and automation to solve your business problems.