- Word clouds
- Distributed flavour
After implementation of SAS Visual Analytics, we overcome the prediction of future business and we were able to plan the future business, who are our premium customers, where we can cross-sell our products more, etc.
If there could be 100% feasibility of forecasting feature available then we could be more accurate as it's currently 95%.
I've used it for one year.
There have been no issues deploying it.
We have issues with the non-distributed flavour sometimes, as the LASR performs slow due to huge amount of data, so we need to fix it by performing a server restart.
Though the LASR has in-memory feature, but due to heavy load and computed columns in the project logic, sometimes the is performance slow, but at the same time in Distributed flavour with Hadoop the performance is always optimum.
There have been no issues scaling it.
I used SAS BI as well. Infomap Studio, Web Report Studio, and Information Delivery Portals(IDP), which does almost the same as this does, but much slower, and there is no dynamic look and feel, and no graphs or charts options available. Also, in SAS Visual Analytics we can prepare data exploration and reports in one platform.
It was a completely new installation on a Unix server, so it was interesting and challenging, but if you have the installation documentation such the prerequisites and post-validation then you can replicate the installation in different environments.
We have implemented both ways at client side as well as in-house, but I recommend that if you implement it in-house then you should have lot of projects with you otherwise it could cost you a lot for the yearly license.
The next release of SAS Visual Analytics, v10, will be cloud based, so pricing and licensing will cost you less.
It doesn't matter if the data is live or historical, if you have visualization and analytics skills, then Visual Analytics is the best way to project the visualization of data to vendors and clients. Because the one time you invest and gain throughout the year, there are so many other tools suchas R, Matlab, Tableau, amd Qlikview available on the market, but there are certain limitations in those.
I would suggest that you always keep it simple and keep documenting what you do. Before working for any vendor, it is always better to train a certain number of people in-house, so at the client side things will be hassle free.