AI works wonders on your BI, like Organic Milk works wonders on your Health
(note: this is an evolving blog – current as of 2021.1.5 September 2021 release - that may be continuously edited as SAS Viya 4 capabilities are updated or added)
It’s no surprise that as a result of the global COVID-19 pandemic, people have become more health and wellness aware. The most obvious change in consumer behavior is the increased popularity in organic and plant-based ingredients. For example, there are seven ways organic milk works wonders on your health. As I write this posting here about updated and new ways that SAS Visual Analytics (VA) for Viya 4 offers capabilities that fall into the AI-enhanced business intelligence category, it made me re-visit my older post on this topic for earlier releases of SAS VA (BI without AI is like corn flakes without the milk). So too there are seven areas where SAS VA for Viya 4 have been enhanced, and likewise new features have been introduced to work wonders on your business intelligence, reporting, dashboarding and data visualization workflows (aka BI).
For those readers not already familiar with my earlier posting on this topic (or a reminder for those who haven’t read it in a while), artificial intelligence techniques (or augmented analytics) democratize advanced analytics to more users by addressing business problems with automated quantitative insights (what I'm “labeling” as AI within both articles). This is a sequel to that original posting of mine. Any capability outlined in the earlier posting not mentioned throughout below is still available in the newest releases of SAS VA.
New features added since SAS VA for Viya 3.x or 9.4:
(in no particular order)
- Automatic Report Review
SAS VA analyzes your reports and presents information in the Report Review pane that can help you improve their performance. The Report Review pane provides you with a description of each issue and offers suggestions for improvements that you can make to your report or to the data that your report uses. For a detailed write-up on this new feature, please click here. Likewise, for a video demonstration recording of this new feature, please click here. The pane also checks for accessibility issues, which can impact users with color-deficient vision (color blindness) or other visual impairments.
- Accessibility
More specific to Accessibility issues, in addition to Report Review, there have been several additions to SAS VA including (but not limited to):
- Text objects enable you to mark text as a heading. Marking headings can improve accessibility for users of screen reader software.
- You can assign a text label to measure-based display rules. The label is displayed in data tips and on the Display Rules pane when viewing the report. The label is also announced by screen reader software.
- Reference lines in graphs are now accessible to screen readers.
For more detailed guidance on creating Accessible reports using SAS VA and the capabilities offered, please consult this documentation link.
- Logarithmic scale
A new Logarithmic scale option is available for scatter plots, time series plots, line charts, forecasting objects, and numeric series plots. There are two main reasons to use logarithmic scales in charts and graphs. The first is to respond to skewness towards large values (i.e., cases in which one or a few points are much larger than the bulk of the data, e.g. if you wanted to compare net worth of individuals worth $40,000 and $800,000,000). The second is to show percent change or multiplicative factors.
- Outlier Insights & Detection
The new Insights window identifies possible outlier values in the data for the objects in your report. If outlier values are detected, then the lightbulb (aka insights were found) icon appears in the report toolbar. Likewise, measures in the Data pane are automatically checked for potential outlier values. The Outlier icon appears beside measures that might have outlier values. Outlier detection can be disabled by the Disable automatic outlier detection for the Data pane user setting and the disableAutomaticOutlierDetection configuration property.
- Automatic Value Grouping
When you Create a Custom Category that is based on a measure, you can generate the value intervals automatically by clicking Generate groups in the New Custom Category window. You can also generate a custom category for a measure by right-clicking a measure in the Data pane and selecting New custom category. You can specify the number of value groups to create and the prefix to use for the names of the generated value groups.
- Auto Save in a multi-user / collaborative environment
If you are a report designer and you use the autosave feature for reports, the message that indicates that you are working on an older version of the report can be automatically identified by SAS VA if your report has been more recently saved by another user. If you choose to open the updated report, then your autosaved report is discarded. However, if you choose to use the autosaved report, SAS Visual Analytics considers that report to be new and gives it the same name as the initial report.
- Conversational AI
SAS Visual Analytics for Viya 4 includes the ability to create and configure conversational bots to help users with their tasks. Each bot is an autonomous program that can interact with users and respond to questions or provide information for a given topic area. SAS Conversation Designer in SAS Viya uses natural language processing along with a collection of intents, utterances, entities and dialogues to process data and provide information to a user. A bot user does not have to present their questions in a special format, but can simply ask questions naturally the way they would pose them to a human. In the context of SAS Visual Analytics integration, you can display information from a report to embed within your bot. More information on SAS bot development can be found here. And more information on SAS bot publishing can be found here.
Existing features enhanced compared to SAS VA for Viya 3.x or 9.4:
(in no particular order)
Note: because these are existing feature updates, it is assumed the reader already knows how these capabilities facilitate AI-enhanced business intelligence. If you are new to SAS Visual Analytics in general please consult the various resource links throughout below after reviewing this original write-up on this topic.
- Automated Explanation (including Explain Data) & Automated Prediction
* The automated explanation object now displays geographic maps when the response variable is a measure and the selected underlying factor is a geography.
* The automated explanation & prediction objects support the following additional languages: Japanese, Simplified Chinese, Traditional Chinese, Korean, Thai, French, German, Italian, Polish, Spanish, Danish, Dutch, Norwegian, and Swedish.
* SAS VA now displays a notification when automated explanation objects in the background are finished loading results.
* Automated explanation objects identify extreme values (possible outliers) in the characteristics summary. The full list of values is displayed in the Anomalies tab of the details table.
* You can explain the characteristics of data items that are assigned to any object. This feature enables consumers of a report to quickly access the same summary that is provided by the automated explanation object.
* Automated explanation objects list the most important factors in the summary description.
* You can explain characteristics of members of hierarchies that are assigned to an object.
* The automatic explanation object includes a forecast of future values when you assign the new Time data role.
* In the Automated explanation object, groups are no longer calculated and displayed by default. You can select the Show high and low groups option to calculate groups.
* For objects that contain multiple visual elements (for example, automated explanation, automated prediction, and advanced analytical objects), the Displayed visuals option enables you to select the elements that are displayed.
- Object Templates for Analytical Objects
Although the ability to create basic object templates is not new (or page templates for that matter), the new Object Templates for Analytical Objects should enable report developers to think of many new use cases for re-use and re-purpose of analytical insightful dashboard designs that now more integrate with SAS Visual Statistics and SAS Visual Data Mining and Machine Learning. Recall this older posting by the same author on the importance of analytical insightful dashboards.
* Decision trees have a new High cardinality predictors option that enables a greater cardinality limit for the categorical predictor values at each node.
* When you create a new partition data item, you can stratify based on a category variable.
* When you register a model (not just Decision Trees), you can now specify the model repository where the model is stored.
* Decision trees can now display validation statistics in each node and in the details table.
* The Interactive Decision Tree mode is improved with a new Split Node window. This window enables you to better specify the split variable, number of branches, split points, and how missing values are treated. The Split Node window has been enhanced to enable better control over the number of splits and how they are created. You can now prune and split an interior node in the same action.
* The process for splitting a decision tree has been enhanced.
* Generated text is now localized in the following objects: forecasting, network analysis, path analysis, and text topics. If SAS Visual Statistics is not licensed at your site, then the basic decision tree also has localized generated text.
* Options for the network analysis object have changed. The new Network structure option and Link type option specify the structure and link directionality for the network.
* The stress centrality metric for network analysis objects is no longer available. For reports that you promote from previous releases, the stress centrality metric is changed to betweenness centrality.
* The calculation of network metrics for the network analysis object now uses a different SAS Cloud Analytic Services (CAS) action set (the network action set). Some results might be different from previous releases.
Example use cases of what is now possible can be found here.
* Displaying the earliest or latest value for the selected date or datetime item is now possible.
* You can now assign a Date format or a Datetime format to a calculated data item.
* Date and Datetime values can now be used as a number in calculated expressions.
* The properties list in the Data pane now includes a read-only preview of the expression for calculated data items. Only five lines of the expression are displayed. You can click Edit beside the expression in the list of properties to open the Edit Calculated Item window.
* The expression editor window for calculated items and advanced filters has been enhanced with a new layout.
* A filter field is now available when you select data items to create new filters, ranks, and display rules.
* A new search field in the Options pane enables you to find specific options by using label names.
* You can now display a search field in the list control. The search field enables report users who are designing or editing reports to find specific values that they want to select or clear in the list.
* The Add Data Items pop-up window has a new Select all check box.
* You can edit the calculated expression for geography data items that are based on a calculated item.
For a Getting Started with Geo Analytics in SAS VA, please consult the linked recorded webinar and material. Enhanced features that have been released for Viya 4 since then include:
* For geo coordinate and geo region-coordinate objects, the Enable data label thinning option hides data labels as needed to avoid overlapping text.
* Geography items have a predefined role for three-letter ISO codes.
* For geo coordinate objects and geo region-coordinate objects, new options enable you to control the size of markers and bubbles.
* You can disable zoom and pan controls in geomap objects.
* The geo line and geo line-coordinate objects enable you to display line segments on a map. You can use geo line maps and geo line-coordinate maps to display linear geographical features such as roads, railroads, pipelines, bus routes, and so on.
* The geographical lookup values for SAS Visual Analytics have been updated with the following changes: 1) The country name for Macedonia is changed to North Macedonia; 2) Localization support for country subregions is improved. For example, the subregions of India now support names in Devanagari script in addition to Latin characters.
* The maxRowsLookup.geoLine configuration property specifies the maximum number of result rows for polylines in a geo line map that is in SAS Visual Analytics reports. For more information, see Configuration Properties: Report Data Service.
* The OpenStreetMap High Contrast map background is available for geographic maps. The High Contrast map is optimized to provide the best usability for users with low vision. If you set the map background to Automatic, then the High Contrast map is automatically selected when the high-contrast theme is displayed.
* The interface for Create a Geography Data Item By Using Custom Polygonal Shapes, Lines, or Coordinates from a Geograph... has been enhanced.
* Single Sign On (SSO) authentication to an Esri Server for Data Driven Content or SAS Viya jobs is on the product roadmap for an upcoming release.
Although this was technically doable since the SAS VA 8.x releases courtesy of Web Content or Data Driven Content, The new job content object enables you to natively connect to and display the output from SAS Viya jobs. You can use jobs to access the functionality of SAS code. And as was mentioned in the previous version of this posting, although not everything one can incorporate with custom built web content, data driven content or SAS Viya Jobs classifies as true AI-enhanced business intelligence, there are examples that can be developed.
CONCLUSION
Just as consumers are looking for reasons to eat healthier and they want more information about where their food comes from and how it was produced, report “consumers” are looking for reasons to automate their data analysis and they want more information and assistance from the software what is contained in their data and how it should be interpreted. Any AI-enhanced business intelligence capability you can introduce to your user base, the more you can reduce time-to-insight and increase decision-making speed. SAS Visual Analytics continues to add or enhance these descriptive, predictive and prescriptive analytical capabilities so that citizen data scientists (in a true self-service manner!) can really contribute to their business success.
