Table of contents for IBM World of Watson 2016
- #IBMWoW: Advanced Analytics in the Cognitive Business
- #IBMWoW: Simplifying and Scaling Cognitive Computing with Watson
- #IBMWoW: Welcome to the World of Watson
- #IBMWoW: Cloud and Data: The Foundation for Cognitive Business
- #IBMWoW: Watson Machine Learning as a Service
- #IBMWoW: Evolution of the IBM Cloud Platform
- #IBMWoW: Ginni Rometty Keynote: Building a Cognitive Business
I am attending IBM’s World of Watson and will be blogging as much as I can. First up is a two-part session on Advanced Analytics and how you can put advanced analytics at the heart of a Cognitive strategy. Paul Zikopoulos kicked things off talking about the potential for data to transform business and the huge amount of data being generated 24×7. Yet, he points out, 24×7 real-time decisioning not so much. The IoT, of course, will explode even the current levels of data and that, he says, leads you to need Cognitive capabilities.
Self service, he says, has largely failed in most organizations. People are not really self-serving, they are still using other people’s creations – the new tools made it easier to build things but did not really change the paradigm. Self service though requires data that can be trusted, tools that allow business and IT to collaborate, ways to check and manage bias and much more. This complexity is what led IBM to create Watson Analytics. And at the end of the day it’s all about improving the quality of decision-making.
Mark Altshuller joined him to discuss the vision for analytics at IBM:
- Expand the user base to include citizen analysts
- Rethink the UI around IBM’s design thinking principles across products
- Make it easier to connect and use data
- Smarter self service
He presented the data to insight lifecycle – Operational Reports to Data Discovery/Predictive Analytics to Enhance/Operationalize to Smarter Decisions and repeat. This is going to frame the discussion in the session, he says. I like the focus on decision-making but I prefer to start with the decision and work back to the data 🙂
A short video of the new capabilities as part of the continuous delivery of IBM’s analytics platform followed with lots of new UI and functions. The new UI is also becoming more embeddable and customizable with new visualization capabilities (shared between Cognos and Watson Analytics), geospatial mapping and data management capabilities.
The new geospatial mapping capabilities are focused on a wide range of geospatial problems, up to and including displaying real-time data streams on maps (demonstrated by Mapbox, one of the new partners), analyzing tweets to see who is a tourist and how they move relative to locals etc. Indeed processing some data, like phone operating systems, can show the map without the map and show divides like gentrification of city based on Apple v Android. More accurate processing allows the lanes on roads to be identified and analyzed separately.
Data discovery was introduced in Watson Analytics and the usage of this original interface was instrumented and analyzed, allowing IBM to simplify and streamline the UI in more recent versions. This usage also identified some very common and strong use cases to be identified to allow these to be made much easier also. Most recently, Cognos data packages can be included in Watson Analytics. Moving forward, Watson Analytics is focused on adding new algorithms, new uses for Cogntive.
For operationalization, Ritika Gunar came up to discuss IBM’s commitment to Apace Spark as the “Analytics Operating System”. This original commitment has led IBM to become a major Apache contributor. The new Analytics IDE – the Data Science Experience – was next and has been widely extended with partners. This IDE is focused on three things:
- Built-in learning because things are changing fast
- Create using open source or commercial add-in
- Collaborate around the data and analytics
Very familiar modeling workflow UI has been used to make it easy for people to use the machine learning capabilities in the new environment. The environment also has versioning, collaboration tools, notebooks that include multiple environments and job scheduling. SPSS models can be integrated and new models can be built in a drag and drop canvas. All on Apache Spark. This can be integrated with Watson ML for deployment.
The CIO of Ameritas came up to tell a customer story. Their focus was on self-service (no data scientists – business and actuarial people), cloud for agility and a vendor with longevity. They use Watson Analytics, SPSS Modeler on the Cloud and Cognos Analytics on the Cloud. And they are watching Cognitive a lot as they see the application of Cognitive to analytics as a critical next step.
Alistair Rennie introduced some of IBM’s planning and forecasting capabilities. A Cognitive business, he says, is a thinking and agile business. It’s not just about improving decision-making but being able to share and extend this drives agility. In particular he says, this really changes the performance management and planning environment. The platform increasingly connects sales, operations and finance, delivering capabilities for planning and forecasting that are more integrated.
Bill Guilmart introduced some clients: Zions Bancorporation came up to discuss their use of the integrated planning and compensation management tools as well as Watson Analytics. They particular liked the more integrated, more real-time/on-demand approach as well as better explanations. GCI Corporation (Alaskan Telecommunications) also came up and talked about more integrated capital investment management, project tracking and more. Bill gave a quick run through of the new collaboration and visualization capabilities in the planning and management capabilities.