Our blog

15.12.2014 21:18 | Our blog

Test Driving IBM's Watson Analytics Beta

Source: forbes.com
Tags ibm, big data
IBM IBM +2.07%’s Watson Analytics beta was made available to the public last week, and shows potential to break down the barriers that hold back analytics adoption in enterprises.
.
22,000 people signed up for the program since the launch on September 16th giving IBM has a strong base to build on for testing future enhancements. Watson Analytics is built on IBM’s SoftLayer cloud infrastructure, and the company is investing over $1B in Watson Discovery Advisor, IBM Watson Explorer and IBM Watson Analytics over the next few years.
.
Relying on a freemium model to gain broad adoption and use, IBM has made the application available via browser, Android and Apple . Advanced tiers of services will eventually be made available over time for an additional charge. IBM is not and has never been a client or mine, and I have never been an employee.
.
Breaking Down The Barriers That Hold Back Analytics Adoption
.
Watson Analytics’ three main areas are Explore, Predict and Assemble, each designed with the needs and skill sets of business analysts and decision makers in mind. The focus is on how to break down the barriers that exist today between decision makers and the data they need to get work done. Watson Analytics’ natural language-based cognitive service interprets and responds to questions quickly and with vary degrees of accuracy depending on the data set being used.
.
Also included are predictive and analytics tools and streamlined data preparation and coding. The Assemble section of the application has a series of dashboard and infographic templates that can be quickly applied to any data set as well.
.
Key take-aways from test driving IBM’s Watson Analytics beta include:
.
  • Data loading and data quality are intuitive, supported with guidance on how to improve data sets’ quality quickly. In this beta, IBM is supporting .CSV and Microsoft MSFT +0.34% Excel files of 50 columns or fewer and smaller than 10MB. IBM has provided an excellent guide to data load and quality titled Introduction to IBM Watson Analytics Data Loading and Data Quality. Watson Analytics provides feedback on the quality of data sets, which accelerates learning how to get better results quickly from analysis. For purposes of the test I used Data.gov, which has 133,731 free data sets available. The 2012 E-commerce Multi-sector Data Tables (Released May 22, 2014) from the U.S. Census was used for the test drive. Uploading the .CSV file takes seconds and the data set appears on the introductory screen of Watson Analytics at the far left selection as shown below:
.
.
  • Selecting Explore delivers several recommendations for initial analysis rated by their relative relevancy. Based on the data set created, Watson Analytics provided the following ten initial recommendations for analysis. Typing in a query leads to the most relevant recommendations based on the constraints of the data set and the linguistic interpretation of the query. This area of Watson Analytics is useful for drilling down into data using questions instead of a series of commands or menu selections.
.
  • Transitioning from natural language query to visualization happens in just one step, with Explore providing a series of options for presenting data. Visualization types supported include bar, grid, area, bubble, line, pie, tree map and categorical charts. The following figure shows a tree map of 2012 e-commerce shipments in the manufacturing sector, with a scrolling ribbon across the top of key analytics from the data set being used:
.
.
  • Aggregating analysis on a single screen to create dashboards can be driven entirely by queries. Using the natural language-based cognitive service, it’s possible to get exactly the relevant series of data sets included in a dashboard as is shown below. This was entirely created using just two steps after a series of text queries for e-commerce data.
.
.
Bottom line: The latest beta of Watson Analytics shows IBM is making solid gains on enabling business analysts and decision makers with cloud-based apps that can streamline their daily jobs.
Site development SeaDesign O2