September 14, 2017

The MineThatData Forecasting Challenge!!

Ok, time for some fun, if you are an analytics guru!

I have a dataset for you ... it contains annual sales on a rolling twelve month basis for the past 83 months (month = 00 through month = 79). You can click here to download the Excel Spreadsheet.

Here are the variables in the dataset.
  • Month = 00 to Month = 79.
  • Annual Sales:  Total.
  • Annual Sales:  Call Center.
  • Annual Sales:  Website.
  • Annual Sales:  Email Marketing.
  • Annual Sales:  Search (Paid and Natural).
  • Annual Sales:  From Items New In The Past Year.
  • Annual Sales:  From Existing Items.
  • Annual Sales:  From The Primary Merchandise Category This Company Sells.
  • Annual Sales:  From All Other Merchandise Categories.
  • Annual Sales:  From Items $0.01 - $9.99.
  • Annual Sales:  From Items $10.00 - $19.99.
  • Annual Sales:  From Items $20.00 - $29.99.
  • Annual Sales:  From Items $30.00 - $49.99.
  • Annual Sales:  From Items $50.00 - $99.99.
  • Annual Sales:  From Items $100.00 - $249.99.
  • Annual Sales:  From Items $250.00+.
A few key tidbits:
  • Total = Call Center + Website + Email Marketing + Search.
  • Total = New Items + Existing Items.
  • Total = Primary Merchandise Category + All Other Merchandise Categories.
  • Total = Items $0.01 - $9.99 + Items $10.00 - $19.99 + Items $20.00 - $29.99 + Items $30.00 - $49.99 + Items $50.00 - $99.99 + Items $100.00 - $249.99 + Items $250.00+.
Here's the Forecasting Challenge:
  • Given the data in the dataset, predict Annual Sales in Month = 80, Month = 81, and Month = 82.
Forecasting Challenge Extra Credit:
  • Given the data in the dataset, predict Annual Sales for each variable in the dataset for Month = 80.
  • Then, use the data in the dataset and the prediction for Month = 80 to predict Annual Sales for each variable in the dataset for Month = 81.
  • Then, use the data in the dataset and the predictions for Month = 80 and Month = 81 to predict Annual Sales for each variable in the dataset for Month = 82.
The person / team / company that minimizes squared error for Annual Sales - Total in the Forecasting Challenge will be deemed the winner. Yes, I have the actual values for months 80/81/82.

The person / team / company that minimizes squared error for Annual Sales - Total in the Forecasting Challenge Extra Credit will also be deemed a winner. Yes, I have the actual values for all variables for months 80/81/82.

The person / team / company that wins either challenge will have their results published on this Blog (2,500 subscribers), on Twitter (6,600 followers), and on my Podcast (71,000 downloads). So yes, your answer will receive some exposure!  I will also send a free copy of Hillstrom's Fix It to the winning person / team / company.

Your solution must be written up in white paper format - fully explaining the methodology you use, and fully sharing the equations or math used to make the prediction. If you don't provide a white paper publicly outlining/sharing your methodology/results/predictions, you will not be eligible to win.

Cost to Enter = Free!

Cost to Submit a Potentially Winning Solution = Free!

Due Date:  Your solution is due no later than 11:59pm PDT on October 31. Late submissions will not be accepted.

Ok, who is ready to accept The MineThatData Forecasting Challenge??!!

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

Items That Appear In Multi-Item Orders

In a typical Life Stage Analysis within a Merchandise Dynamics project, it is common to see exaggerated trends when comparing first-time buy...