Introduction Scholastic’s Classroom Magazine business sells magazine subscriptions to individual teachers in grade schools (elementary, middle, high) for use in their classrooms. There are 26 different magazine titles available, each focused on a different grade and subject matter (reading, math, etc.). One subscription to one of these magazines contains enough copies to distribute to an entire class (usually 15-35 students). Sometimes a school will have a central coordinator for magazine subscriptions, and sometimes the teachers place their orders individually. These subscriptions are paid for in two main ways - the teacher can use their own money, or the subscription can be paid for by the school district in which the teacher’s school resides. The magazine business co-exists with three other Scholastic businesses doing business in grade schools - a catalog business, an on-site event business, and a B2B business selling educational solutions. Request from Magazine Business While annual revenue increased from 2017 to 2019, due to a price increase in 2017, we saw at least a 1% decline annually in subscriptions from 2017 to 2019. The Magazine business’ request to the Data Science team is to conduct analysis and give a recommendation for how Magazines can increase the number of subscriptions via marketing.
Questions:
- Conduct an exploratory data analysis to characterize buildings that contribute to the most loss. Please summarize your findings. (Note: The data provided is limited to elementary schools - i.e. schools containing grades up to 6th grade)
- In order to fulfil the Magazine business’ request, the Data Science team is tasked with building a predictive model to make data-driven marketing decisions. Using the data provided, your task is to build a model to identify which school buildings Magazines should target in order to increase subscriptions. a. You may choose to build one of the following models that could lead to impactful recommendations. i. an attrition model, identifying buildings that are at risk for reducing or not renewing their subscriptions, ii. a model uncovering the features of high-performing schools and potential upselling opportunities, iii. or you may choose to develop a solution of your own that could bring values to business. b. Prepare a summary of your modeling work and results, as if you were explaining to the rest of the Data Science team. Please be specific on all the steps you take to build a final model including exploratory data analysis, data pre-processing, feature construction and model performance evaluation.
- Based on the learnings from the model built in question #2, what are the recommendations you would make to the Magazine business? Please prepare a few slides summarizing your findings and recommendations, keeping in mind how you would explain the results to a non-technical audience. You should be prepared to present your analyses, findings and recommendations during your Super Day interview.