Table of contents for IBM Watson Analytics Cloud
Dave Laverty kicked off IBM’s big Watson Analytics announcement event in New York. The focus today, he says, is the reinvention of the analytic experience with a degree of simplicity that puts analytic power in the hands of everyone. Bob Picciano joined him to make the key announcements.
Back in January IBM announced the Watson business unit, a vertically integrated group focused on the use of cognitive computing to solve tough problems. This group, he says, has made great strides with the underlying technology and today’s announcement is about using this technology more broadly and making it accessible to a broad community of business analysts.
IBM sees three key trends
- The use of data for competitive advantage
- Cloud enabling new business models
- People-centric engagement is changing the way companies and organizations interact with people
The data, he says, is the “what” that underpins the “why” of engagement and cloud is the “how”. IBM likes to talk about Systems of Engagement to pair with Systems of Record and is increasingly also talking about Systems of Insight. Systems of Insight drive analytic insight into every decision and interaction in every process. These systems are driven from the data in systems of record but increasingly also tap into the new information captured in systems of engagement.
Leveraging analytics in this way is still a challenge, however, with 38% having a limited understanding of how to use analytics, 80% of time being spent not on analytics themselves but on data preparation for those analytics and 24% finding it hard to get the data they need at all.
The new capability is designed to focus on problem solving, answering the questions that matter to companies. Focusing not just on visualizing the data but on analyzing it to actually answer questions. They want to put analytics in the hands of “everyone”, making data access and refinement easier and then deliver this through the cloud.
Three announcements therefore:
- Watson Analytics, a unified self-service analytic platform for business users delivered in the cloud
- Data Refinery, a set of data services designed to make data useful and relevant delivered in the cloud.
- Analytic Warehouse, an agile warehousing solution for fast deployment of analytics in the cloud.
Marc Altshuller and Neil Whitney came on stage to give a little more detail on IBM Watson Analytics and how it can be used to solve real-world problems. Even for simple analytics problems there are multiple steps involving multiple people with everything from data access to data preparation, reporting to advanced analytics. Several elements are key to the new product:
- It’s self-service and cloud-based
- It uses natural language based on the Watson platform
- It’s designed to use any kind of data
- It’s designed to support the analytic journey for people of different skill levels with teasers and thumbnail summaries.
The demo had some interesting features:
- Basic management interface
- Story-based templates and samples such as “find patterns in wins and losses” as an alternative to stating with data.
- Many different datasets can be accessed either by uploading or connecting to data sources so it should be easy to pull in a wide variety of data
- The interface shows that data is trusted if the governance team has flagged it as such. Data quality scores are also generated for columns.
- Nice collaborative features allow team members to discuss, ask each other questions etc.
- Many workbooks can be created and each has a visual thumbnail, using visualization technology built into the platform
- Starting a new workbook
- The cognitive capabilities are applied to data as soon as it is loaded up, suggesting interesting columns in the data.
- It also suggests starting points with a clue for the potential value of a given starting point given using different sizes.
- The tool suggests potential analysis approaches and attributes as well as flagging data that needs clean up first.
- Each column is presented with summary data, histograms etc.
- The engine suggests potential data clean ups – like merging USA with United States, replacing blanks etc. Each can be reviewed individually or the user can simply say “clean it up” and apply all of them.
- The tool uses machine learning to suggest predictors and can show how the predicted values match actual data
- Visualize these predictors in a compact yet effective way that is interactive and can show impact of suggested predictors as well as those selected by a user.
- It provides some nice visual tools, word clouds using size to show importance of predictors as well as a clear distinction between places where the prediction is strong and where it is weaker in a decision tree for instance.
- The interactive tools can be combined with color coding, regional maps etc.
- Nice wizard-based interface generates presentation, infographics, based on the work being done
- All the favorites and interactions that went into the analysis are remembered and used to drive the infographic with the engine selecting appropriate templates for the various elements.
- Users of these can select visualizations and “pin” them into their dashboard environment.
Go to market is driven by a freemium model with a view to get this in the hands of users as quickly as possible. Users will be able to do useful things in the free version but the intent will obviously be to have uses move to paid versions over time.
Very impressive overall. Get access at www.ibm.com/analytics/watson-analytics/