SAC & GCP: SAC Forecasting with Google Cloud Services
- Google Cloud Platform, SAP Analytics Cloud
- Forecasting, Google Cloud, sac, sap analytics cloud
- 4 min reading time
Gary Lude
The introduction of the SAC Export & Import API opens up new possibilities for overcoming the previous forecasting limitations of SAC. With the help of the Export API, data from a SAC planning model can be easily transferred to the Google Cloud Platform. All GCP services are then available in the Google Cloud Platform, including the option to implement individual forecasting using R models, for example.
Table of contents
Google Cloud Services and SAC Forecasting at a glance
The SAP Analytics Cloud (SAC) already provides native forecasting functionalities for predicting planning data. However, these are often inadequate for individual requirements or deliver inappropriate results and also do not offer the option of including additional external data sources in the forecasts. Nevertheless, more and more companies want to automate their forecasting process as much as possible. This makes it possible to quickly and easily gain insights into the development of the coming periods in order to make the best possible strategic decisions.
The introduction of the SAC Export & Import API opens up new possibilities for overcoming the existing forecasting limitations of SAC. With the help of the Export API, the data from a SAC planning model can be read very easily into the Google Cloud Platform (GCP). All GCP services are then available in the Google Cloud Platform. This also includes the option of implementing individual forecasting, e.g. with the help of R models, which covers practically all possible requirements and ideas. Once the forecasting process has been completed in the Google Cloud, the forecast values are written back into the original SAC planning model via the import API and can be revised manually or used directly for reporting.
How does forecasting work?
A web application is provided for the end user via GCP Cloud Run, which enables interaction with the forecasting process. The forecast can be started here, the current status can be queried and the forecast values can be imported into any target version of the SAC.
When the forecast is started, the data from the SAC model selected via dropdown is exported using a Google Cloud Function via SAC Export API and sent to Google Cloud Pub/Sub as suitable packages for the forecast. Each package starts an instance of Google Cloud Run, which contains both a processing level with Python and an underlying forecast level with R. Instead of forecasting using the R model, any other programming language would also be conceivable and possible.
As the technical basis of Cloud Run is docker images, it is possible to use a wide range of programming languages and environments.
As a forecast can be quite complex and a run can take several minutes, the forecasts of the forecast run are first written to Google Cloud BigQuery, GCP's data warehouse, by the respective Cloud Run instance. The forecast is created in the background and the user does not have to remain logged into the online session in front of the computer. However, they can query the status of the current run at any time within the web application if they are interested. The status is determined technically by comparing outstanding and already executed forecast packages within a cloud function.
Once the forecast has been completed, the final predicted values can be imported into the SAC. A cloud function is also used for this, which now uses the SAC Import API to write the data to the stored SAC planning model. A short time later, the forecast data is then available in the SAC for further use.
What are the advantages of this solution?
Features of this approach
Creation of an individual forecast
Any forecasting scenario can be implemented with our solution, allowing us to offer our customers a high degree of flexibility without restrictions. Together with the customer, we evaluate the optimal process and the services required to realize the customer's ideas. The customer can also decide how users can or should interact with the forecasting. There is also the option of fully automated forecasts, which are always executed automatically at a certain point in time. Our principle is: give free rein to your ideas and we will make sure they become reality.
Customized web application
It must be possible to cover the individual needs of each company in order to enable easy integration into the working environment. Based on this, we have opted for an application that can be easily adapted to individual requirements. The focus is on the interaction possibilities by means of forecasting, whereby there are no limits to further modifications. The use of Python Flask enables a high degree of flexibility, even for the fulfillment of complex requirements.
Using the SAC Export & Import API
The natively available SAC Export & Import API offers an interface provided by SAP for communication with SAC planning models. We have integrated this technology into our solution to ensure a stable, secure and functional export and import option. Furthermore, the interface is actively maintained and constantly developed by SAP.
What opportunities are you missing out on without this solution?
Are you having difficulties implementing your forecasting requirements in the SAP Analytics Cloud? Can't find a way to implement your forecasts individually? If so, you are missing out on the full potential of your data. This is exactly where our solution approach comes into play.
With Google Cloud Services as the basis for your forecasting process, you can convert your business cases into a stable and scalable solution with us. In addition, your analysts can interact with an intuitive user interface if required and control the forecasting individually.
Without this customizable solution, you may not be able to implement your complex forecasting requirements as desired and lose important insights into the future development of your business. We are happy to support you in turning your ideas into reality.
Know more?
Would you like to delve deeper into this topic? Then we look forward to talking to you personally about the possibilities of the Google Cloud Platform (GCP).
