Wawa HCL Hackathon 2019

Problem Statement:

• Find a way to predict future sales at a store level for Product X.
• Produce a BI product that allows a user to easily visualize and explore the implied impact of each factor by store by key category.
• Address a given product X

Requirements:

• Any generally available dataset from Wawa or an external source may be used. Some examples to consider are weather, labor, and marketing campaigns.
• Given a sales dataset, the work product should produce a forecast of future sales.
• The work product should also produce a summary of the implied impact of each factor on sales during the forecast period.
  • This summary should be available by store by category, so that Robert can easily tell what’s causing his sales to vary.
  • Clear visualizations should be available so that Robert can spend his time managing the store rather than staring at spreadsheets. He should be able to easily explore the results visually.

Data Provided:

• A sales dataset will be provided.
• A list of MD stores open at least 3 years will be provided.

Click here to download

Success Parameters:

• The quality of the predictions will be judged by averaging the Mean Average Percent Error of each store/category prediction using the out-of-time sample that has been withheld.
  • A 1 year period will be withheld for this purpose. This dataset will be provided to each group near the end so that they can present visualizations using this period. No group shall be permitted to use the holdout data for any other purpose than making final predictions and feeding the results to the BI product.
• The quality of the BI product will be judged by the ease with which a user can visualize the implied impact of multiple factors, as well as how easily the user can explore the results interactively in novel ways.