Having had the opportunity to have some great conversations with a number of people (at SAP, with customers and internally) on the SAP Predictive Analytics Suite there’s been a topic which has cropped up a few times – does Predictive Analytics sit in with general BI (which resides in the hands of the wider business in terms of consumption) or with a specialist team of predictive experts (aka data scientists etc.)
Ultimately the answer is that it sits with both. The SAP Predictive Suite offer value to both business users and predictive analytics experts within an organisation. One of the really compelling elements of the SAP Predictive Analytics suite is that it offers two core components – Automated Analytics and Expert Analytics, meaning multiple modes relevant to the type of user.
The automated analytics element of SAP’s offering means that business users, or the BI team, can quickly and easily build predictive models as well as deploy and manage them. This means we can deliver value adding predictive models in a rapid and agile manner – the tool is doing the hard work for you.
The concept of automated analytics complements the predictive experts you may have within the organisation, this individual or team can focus on the business critical models and let the business run with the more routine, high volume predictive modelling. They can use the Expert Analytics tool to build complex data integration and preparation steps before passing data through a model (either a delivered, HANA or a custom algorithm).
For me, Predictive Analytics should be seen as part of BI (and vice versa) as they are both so closely linked. Whether you are training or consuming a predictive model, you need the data modelled in a particular way. Data modelling is just as important part of predictive as it is BI. There’s also the ability to use BI to analyse the results of the predictive models as part of the model below: