Member-only story
Fine Tune Sales Forecast with Prophet Logistic Growth
This is a tutorial on how to both forecast growth and saturate minimum by using Logistic Growth trend model, by specifying cap and floor, respectively to Prophet Forecast Model to hopefully increase forecast accuracy and reduce Mean Absolute Percentage Error (MAPE). This tutorial complements existing tutorials — both Forecast Sales Using Prophet and Fine Tune Sales Forecast with Prophet Regressors where it uses Logistic Growth trend model, instead of the default linear model for its forecast.
Using Logistic Growth is helpful during the period of uncertainty (e.g., pandemic) where you can use both cap and floor to massage forecast to account for both 25% drop in retail sales in 2020 and 35% increase in retail sales in both 2021 and 2022.
To maintain the brevity of a tutorial, some of the important steps are not referenced in this tutorial. For a complete code walk-through, please refer to the Github repository referenced in the Resource section at the end of this tutorial. Please note that this tutorial is very similar to Fine Tune Sales Forecast with Prophet Regressors with minor differences. This tutorial was created for me as reference material, therefore the tutorial is heavy on how-to, not on algorithms.
Prerequisites
The following are prerequisites for this tutorial:
- Forecast Sales Using Prophet (https://sungkim11.medium.com/forecast-sales-using-prophet-90b6bbd18f8f)
- Fine Tune Sales Forecast with Prophet Regressors (https://sungkim11.medium.com/forecast-sales-using-prophet-with-regressors-b19adf8080ab)
- Data
Data
Four data sources were used for this tutorial where United Stated Census Bureau Monthly Retail Sales is used to forecast future sales and both FRED Data and OECD Data were used as Regressors data.
United States Census Bureau Monthly Retail Sales
United States Census Bureau maintains Monthly Retail Trade Report, from January 1992 to the present. This data was picked to illustrate forecasting because it has extensive historical data with the same monthly frequency. Data is available in Excel spreadsheet format at https://www.census.gov/retail/mrts/www/mrtssales92-present.xls