We need your help extracting insights about our customers. Below is a large dataset with information on customer demographics, the size of their bank balance (1-low to 5-high), account information (type a, b, c...), and the specific channels we use to reach this customer (i, ii, iii...). This is all *real* customer data that we have masked so that we can share it with you.
Our objective is to learn which outreach methods yield the best product portfolio. In other words, what customer interactions are most related to size of portfolio and balance growth?
Five winning students will each receive $2,000 and a 3-day/2-night trip to San Francisco, CA in March 2017 to attend a Wells Fargo Analytics Summit.
Then submit the following deliverables which answer the main question: "What sequence, frequency and combination of channel outreach optimizes product portfolio (both depth and balance)?"
Deliverable 1) Describe your analytic process and technical results– one to two pages
i. Results from your exploration of the data
ii. Description of changes you made to prepare the data for model development (filtering, transformations, etc.) based on this exploration
iii. What model methodologies did you consider
iv. Which did you use for your final submission and how did you make the decision (what metrics did you use)
v. Are there any caveats about using your chosen model(s)
Deliverable 2) Document the code used for your final model(s) – no more than two pages
Deliverable 3) Present results and insights – one to three pages describing what the results from the modeling indicate and how they can help the business (visuals would be very helpful here).
i. What drives growth in accounts and/or balance between month 0 and month 12.
ii. What demographic types, if any, are more likely to increase (or reduce) their number of accounts and/or balance between month 0 and month 12?
iii. What types of accounts, customer interactions, customer events, or Wells Fargo outreach, are more correlated with account and/or balance change?