Revenue Forecasting – Solution Demo
Custom solutions using accelerator templates:
- Balance ease of use with flexibility and robustness is essential for driving adoption in any solution
- Enable custom applications and implementations built on a solid framework
- Purpose-built for your use case, backed by our knowledge library
Forecasting demo walk-through
In this example, we will demonstrate how easy it can be to start leveraging machine learning forecast data into an existing workflow through a custom forecasting application. By adding an application layer on top of machine learning models, finance departments can standardize their approach to producing baseline forecasts through a common portal. Below is an example of a revenue forecasting process.
Upon launch, the application provides an option to upload an Excel file containing the data you would like to leverage to produce a forecast. The application then populates the subsequent fields with the column names from the uploaded Excel document.
Once uploaded, the user may choose the following fields:
- Target column: the field for which you would like to produce a forecast
- Partition column: how the data should be logically separated. This could be a product field, location, or SKU. In this example, multiple partition levels are encoded into one column.
- Date column: the column containing the timestamps of the observation
- Number of horizons: how many periods into the future you would like your forecast to be generated for
- Frequency: the frequency of the data (e.g., daily, weekly, monthly, or quarterly)
- Models: in the advanced case where you would like to expose model options to the end-user, they may choose which models to consider when generating the forecast.
Once the required fields are populated, the user can submit the data for processing. It can take 2 – 10 minutes for results to return, depending on the amount of data and the number of models selected. Once complete, the application provides the path to the output file.
Visualizing the result
By standardizing the results’ forecasting process and output format, organizations can leverage a suite of reporting tools that update when new forecasts or data is generated. In this example, historical data is shown alongside the values generated by the fitted model. In addition, a table of forecast values is provided to assess and download future values quickly.
Working with Neal Analytics
Neal’s modular approach to modernizing finance and flexible engagement model make it an ideal partner to modernize financial data. To learn more about modernizing finance and how Neal can help, please contact us.