Tableau comes with a built in dataset referred to as Superstore. This dataset represents products, sales, and profits from an imaginary company. The dashboard in this example showcases how R and Python extensions can be used together from the same Tableau Workbook.
The R extension used in this workbook uses the outForest package to do outlier detection given a set of sales and profit data. The function returns a logical vector / array indicating whether or not a given observation was identified as an outlier.
This extension is used in the Tableau Workbook to visually identify outliers.
The Python extension used in this workbook uses a simple scikit-learn linear model to predict profit given the following inputs:
- Days to ship actual: How many days did it take an order to ship
- Days to ship scheduled: How many days is it expected to take an order to ship
- Quantity: How many items were in the order
- Sales: How much was the order
- Discount: What was the percentage discount applied to the order
Given this information, the extension returns a predicted profit value. This can be used to compare to actual profit to determine orders and customers that are over / under performing according to profit expectations.
The Tableau workbook included in this example is built off of the default Superstore workbook that is included in Tableau. It contains 3 distinct worksheets:
- CustomerScatter: A view of customers comparing Sales to Profit. The color of each point indicates its outlier status.
- CustomerPerformance: A view of customers comparing Profit to Predicted Profit. The color of each point indicates its outlier status.
- CustomerOverview: View of sales performance broken down by various categories across regions.