You are part of an analytics team that is trying to inform amenities design for an interstate highway by evaluation drivers’ preferences. In other words, what type of amenities should be build along the highway.
The data comes from an experiment where a promotional coupon is offered to a driver. The promotion is based on a type of amenity such as Bar, Carry out & Take away, Coffee House, or Restaurant. If the driver accepts the promotion it can be inferred that their preference is toward the same type of amenity service along the highway.
With the coupon a survey is also conducted. The survey describes different driving scenarios including the destination, current time, weather, passenger, etc., and then ask the driver whether he/she will accept the promotion.
Run a predictive model to assess the likelihood of a driver accepting the promotional coupon. Then, interpret which amenities are more important for your positively predicted class.
In addition, using the team submission file, score 2684 drivers as with a 1 if the accept the coupon and 0 if they do not. This file will be evaluated using accuracy and f-score
This project is competition style, asking teams to build a predictive model, tune it, and score a set of 2685 records. In addition, teams need interpret and provide operations analytics recommendations which should emerge from the data and the results of a predictive modeling.
At the end, teams should complete the following:
Presentation: summary of the business problem, work completed, results, and recommendations
Analytics case: what business problem are you solving (be creative)
Model: a tuned model with performance metrics and results
Scores: predictions actions on a set of new users. Visualize results as a histogram
Model performance and technical approach: 50%
Data review & feature engineering
Review of models (please use a table to summarize attempts)
Selected model and feature of importance
Implications and recommendations
team#_submission.csv (TIP: Don’t change the layout of the submission file, only add your predictions)