Last week I attended a virtual event held by IBM that covered the upcoming release of IBM Cognos – 11.1. The significant changes IBM has introduced into the latest version revamps the existing tool with many new, future-forward features. Below I’ll highlight the key features worth noting.
AI Conversational Assistant
People familiar with IBM Watson Analytics will see that it is no longer a product available for purchase. Leveraging the innovations from Watson Analytics in natural language processing (NLP) against data, we now have what IBM calls the AI conversational assistant in IBM Cognos. It gives the ability to use natural language queries to ask questions against data. This is very useful when exploring your data. It lets you ask questions such as:
What influences the column sale quantity?
Where is country stored in my data?
What is the quality of my user data?
User Centric Data Preparation and Modelling
In 11.1, IBM made available a host of self-service modelling capabilities to give users the ability to join data sources, insert complex calculations, setup determinants and many other features in the past only available within framework manager. This shifts the workload from IT resources into the hands of analysis.
Cautionary comment: This could degrade trust in the BI solution if analysts are unaware about how to work with their data. Take the scenario of incorrect joins, unvalidated external data or incorrect determinants. User-centric data prep and modelling needs to be rolled out with a plan that takes into account data governance.
Here is a helpful video from IBM showcasing the user-centric data preparation and modelling:
Machine Learning Visualization Recommendations
Cognos will recommend visualizations to use when exploring your data. The recommendation engine considers previous visualization choices as an input. This means, if you dislike pesky pie charts and subsequently don’t use them often, overtime they won’t be recommended as frequently or even at all.
Another extremely useful aspect of the ML addition is found when developing new visualizations on your dashboards. A series of potentially related visualizations can be explored on the fly. These options incorporate Cognos’ ability to identify influencing data behind the metric in its recommendation. For example, if you pull in a simple visualization that shows revenue by region, the tool will recommend additional charts based on drivers that heavily influence the main metric such as location, discounts, etc. This can affirm or lead to new data relationship discoveries and allows a data analyst an explorative unbiased approach to building a dashboard.
Natural Language Generation (NLG)
With the demands of wanting deeper insights on data, we are seeing more complex visualizations being leveraged for advanced analytics: sunbursts, circle packing, alluvial diagrams, etc. A data scientist might feel right at home with these charts, but many report consumers will find it intimidating.
In 11.1 on the top right of the visualization is a details option that uses NLG to explain what is being shown to them. This helps a user close the gap in understanding the visualization via text narratives and, therefore, lessen the need to reach out to the analyst for an explanation.
The above are four items I wanted to call-out from what I have seen in the Cognos 11.1 release. There are many more features that make it an exciting release – below is a short list of some of them:
Copy and paste ability of visualizations, etc. from report to dashboard/dashboard to report, etc.
Ability to set your own custom color pallets
Slack integration -> IBM’s emphasis on leveraging existing collaboration tools instead of making a new one
Ability to make inline changes on report object elements including resizing
Docking of the floating context bar – eliminating the times it gets in the way of what you are focusing on
Geospatial charting is improved leveraging GEO JSON via map box
The 11.1 release brings the investments IBM has made in cognitive computing into their flagship analytics product. It has the potential of elevating report consumers into data analysis and reducing IT dependencies and spend.