What is most valuable?
- Ease of use. Can get results quickly.
- Intuitive understanding of different data types. Saves time and effort.
- Descriptive statistics of my dataset. Serves as logical justification for the predictions.
- Presents output in visual form (Predictive Power/Reliability, Key Contributors). Can quickly assess quality of model and gather insights at a glance.
- Alerts regarding suspicious variables. Enables quick fixes to model.
- Dataset modeling functionality. Can be used to quickly enrich the dataset using just a few clicks, as opposed to spending a lot of time considering the different enrichment possibilities and then spending the time to build the dataset accordingly.
- Confusion Matrix. Good for setting the right threshold value.
How has it helped my organization?
My organization is a consulting company that helps its clients in using technology to improve their businesses. This product was the first one that we used to demonstrate capabilities to our customers.
What needs improvement?
- Dataset modeling functionality could be easier to use.
- Neural network interface for examining complex data types like pictures, audio, etc. But I suppose the Leonardo – TensorFlow integration might be taking care of that.
- Data exploration features could be improved, so that instead of only observing the standard deviation or variance of a column, you can also see the skew and kurtosis.
- Data transformation features. I’m not sure if it is possible to normalize the dataset using log transformations and such, but a quick way to do that would be nice.
For how long have I used the solution?
What do I think about the stability of the solution?
Minor bugs but nothing serious. Performs pretty fast, and doesn’t really crash.
What do I think about the scalability of the solution?
How are customer service and technical support?
Which solution did I use previously and why did I switch?
We have also used Microsoft Azure Machine Learning.
Both have their pros and cons. I would say SAP Predictive Analytics is better suited for business users because it hides the complexity of the model (parameter tuning, etc.), whereas Microsoft Azure Machine Learning provides a lot more flexibility for technical professionals to tweak the model (e.g. parameter tuning to improve the Precision, Reliability, or the F1 score, which can be useful depending on the business objective).
How was the initial setup?
Very straightforward. But from what I have observed, people not familiar with machine learning concepts don’t get it right away.
What's my experience with pricing, setup cost, and licensing?
The license fee appears to be prohibitively expensive and overly secretive, leading our clients to opt for cloud-based solutions that only charge for data storage and processing time.
Which other solutions did I evaluate?
What other advice do I have?
Hire Bilot to do it for you the right way.
Which version of this solution are you currently using?
Find out what your peers are saying about SAP, Microsoft, IBM and others in Data Science Platforms. Updated: July 2021.
523,535 professionals have used our research since 2012.