How Do I Get the Most From AI-Enhanced BI With SAS® Visual Analytics? On-Demand Recording and Q&A
Did you miss the Ask the Expert session on getting the most from AI-Enhanced BI with SAS® Visual Analytics for SAS® Viya®? Not to worry, you can catch it on-demand at your leisure.
During this webinar, I demonstrated the AI-enhanced business intelligence features that are baked right into the latest release of SAS Visual Analytics (VA) for SAS Viya. In the recording, you will learn:
- What AI-enhanced BI is and why you need it.
- An overview of 10+ features within SAS Visual Analytics 8.5 for Viya 3.5 that allow AI-enhanced BI.
- Which features are available in older releases of SAS Visual Analytics
- How to determine whether to perform AI or BI first, depending on your data storytelling path.
Here are the questions from the Q&A panel throughout the webinar:
1. How did you run the Automated Explanation report? It looks very powerful, but I missed how you run that in VA 8.5.? Answer:
2. How do some of these AI features like Auto Explanation compare with what a data scientist could do when creating models using Visual Statistics or VDMML?
And there was an additional related question ... How would you compare this to Visual Statistics and VDMML?
Answer: Please see the section paragraph titled What about for the true Data Scientist? within this article that was the basis for this webinar.
3. As a follow-up question to previous Question/Answer directly above ... So then how would you compare Visual Statistics with VDMML?
Answer: Visual Statistics provides objects for descriptive modeling - k-means clustering and "traditional" Machine Learning algorithms like trees, log regression, generalized linear models, etc. Whereas, VDMML adds more modern Machine Learning models like gradient boosting, random forests, deep Neural Nets, Factorization Machines, etc. VDMML also offers different levels of automation courtesy of provided Model Studio.
Relating to SAS Model Studio there was a question asked ... I was in Model Studio and built some models. I went to view results and only found variable importance but can't find the variable impact. For logistic regression model for example, I can't find the odds ratios. Could you tell me how to get that information?
Answer: Odds ratios can be calculated as part of post-fitting using the logisticOddsRatio action which is part of the Regression Action set. This can be accessed through a coding approach (like using PROC CAS in SAS) to program the options for logistic regression in CAS. It is not offered as an option through the logistic regression node in Model Studio as this is typically used more for inferential modelling rather than predictive modelling with large data. You could include this code through a code node in the pipeline.
4. How does SAS VA mapping compare to SAS/GIS from SAS 9.4. Or do we work with SAS/GIS?
And there was an additonal related question ... Do you have any feel how this compares to packages like ESRI?
There was also a third related question ... Can you give/send a detailed demo or post something on your blog for the Geomap capabilities that you demoed?
Answer: SAS VA has built-in geo spatial analysis capabilities that are courtesy of Open Street Map and ESRI.
SAS VA does NOT compete with Open Street Map and ESRI. We provide the data and Visual Analytics functionality. They provide the geo map (i.e. boundary) layers.
For more information, please see the section paragraph titled Demographic data within this article that was the basis for this webinar.
5. There was a mention of "modernization of the SAS Stored Process". Can you provide more details on that?
Answer: Please read this article for those details.
6. Is it possible to do a 'Heat map' in what was being demonstrated?
Answer: Yes. It was not demonstrated because this wouldn't be officially considered AI-Enhanced Business Intelligence by organizations like Gartner 😊
For more information on 'Heat map' functionality, please consult the documentation here.
There was an unrelated question but the answer lends itself to 'Heat map' ... Does the Correlation shown screen out categorical variables?
Answer: Yes. The Correlation Matrix only uses measures. A heat map might be better to represent relationships between categories and measures.
7. Was that SIRI or Alexa in the video you showed?
Answer: The voice assistant feature built into SAS Visual Analytics App for iOS is separate from Siri or Alexa. The voice assistant does not respond to Siri or Alexa commands.
Note: SAS Visual Analytics Product Management are looking for real-life business use cases you could apply voice assistant to your every day work. If you could please share these example use cases, please comment back at the bottom of the following article, thank you and much appreciated.
8. Have you compared the AutoML with trees (and other models) developed by Experts?
Answer: Often times the AutoML algorithms provide better results than first run models by experts. However, Auto Explanation (and similar black box machine learning) only uses a subset of modeling algorithms so a data scientist that can use business knowledge with a broader range of models and advanced data prep can do better. Think of these capabilities being demonstrated here as a starting point for model development.
9. How do you find the independent variable impact (positively or negatively) on the target (aka response or dependent) variable?
Answer: To get more details within any visual that is analytical in nature, click on the 'Maximize' icon as depicted in the following screen capture ...
10. What's the minimal version of SAS VA I need for any of the features covered? Answer:
So then how do I know what version of SAS VA I'm running? Answer: When you're within the report building interface, on the upper right of interface, you can click on Help/About to get that information. An example screen capture from SAS VA 7.5 and SAS VA 8.5 is provided below ...
Recommended Resources
a) This webinar was based on the following article.
b) When you are ready to apply AI enhanced (or any analytical) capabilities to your reports and dashboards, here is a creative way to go about producing them.
c) If you are currently running a version of SAS Visual Analytics that is older than 7.5 and want to explore your upgrade options so you can take full advantage of what was presented during this webinar, please consult this article.
Want more tips? Be sure to subscribe to the Ask the Expert board to receive follow up Q&A, slides and recordings from other SAS Ask the Expert webinars.
Click image to register for webinar
Classroom Training Available!
Select SAS Training centers are offering in-person courses. View upcoming courses for:
