I really value having data preparation, management, analytics, and machine learning all gathered under one roof with SAS Viya. Before using it, I was spending a lot of time reconciling data across different environments. Now, the integrated workflow from data preparation through to model deployment in one environment has streamlined the process. I can trace any figure in a report back to its source, which is a huge help when preparing for internal reviews. Review collected by and hosted on G2.com.

The spatial and geostatistical procedures need refinement before I would trust them unvalidated for location-based modeling, and the documentation around those routines should be considerably more detailed than it currently is. Variogram fitting is sensitive to parameter choices, and the interpolated surfaces tend to smooth over genuine differences between adjacent submarkets when property data clusters unevenly. There's also a lack of diagnostics on fit and prediction uncertainty, which means I have to validate anything for a valuation or investment committee paper in a separate GIS tool. Review collected by and hosted on G2.com.