• Skip to main content

IBM Blueview

Cognos Analytics and all things IBM

  • The Blog
  • Cognos Glossary
  • Cognos Resources
  • About Me
  • Categories
    • Cognos
      • Data Modules
      • Administration
      • Framework Manager
      • Dashboards
    • Opinion
    • Community Spotlight
  • PMsquare
  • Subscribe

Review: Watson Analytics Explore

July 6, 2015 by Ryan Dolley Leave a Comment

Interested in Watson Analytics? IBM Blueview has got you covered. Check out our Watson Analytics first look,  Natural Language Query review and overview of Watson Analytics Professional pricing and features.

Today we’re going to take a gander at the query and visualization component of Watson Analytics, “Explore”. Standalone it is a highly competitive visualization tool that puts Cognos Workspace Advanced to shame; forming like Voltron with Watson Analytics’ other features yields a powerful and unique data slaying super-robot. If you’re curious how Watson Analytics stacks up against Tableau this is the most relevant feature for direct comparison and I can give you a free TLDR by saying: Pretty damn well.

Side note: life doesn’t offer a ton of opportunities to make Voltron jokes, one must seize the lion by the mane… … … sorry!

Launching an Exploration Expedition

Starting the Explore feature requires that you’ve uploaded a data set to Watson Analytics. From there you click on the massive “Explore” tile at the top of the screen and select the set in which you are interested, or you can select the set itself and choose “Explore” from the listed options. This launches a starting point for our exploration:

Starting out on the majestic journey of data discovery
Starting out on the majestic journey of data discovery
  1. The pane at the top is global across Watson Analytics and allows you to save, to quickly swap between open data sets and tools and to see my real name. Ooops.
  2. Here you work the Natural Language magic we covered in our post on said topic, but briefly you type a question in English and will see suggested visualizations that (maybe) provide answers.
  3. Watson Analytics assesses the data set and suggests a number of exploration starting points. You see a brief description of each above an image that tells you what sort of visualization it suggests. In our case “What is the trend of Receptions over Week by Position” is a line graph. Also notice the words in bold – those are column names from the data.
  4. This little button brings up a window that lets you create custom dimensions, add new calculations and fields, and generally mess with the data.
  5. Displayed across this bottom pane are each of the columns from the data set. Adding filters, groupings or calculations by clicking on an item will apply that change across all visualizations in your current “explore” session. Very handy!

If you are a Cognos guru or Tableau fanatic you might be saying to yourself, “This is kinda cool but can I skip all these suggestions and build it myself? As a guru fanatic I don’t need help from Doc Watson!” The answer to that is “No.” You are required to go through this screen to start. I wasn’t thrilled at first either but consider this – Tableau democratized visual storytelling and one-upped Cognos by empowering business experts to create compelling visualizations without IT: IBM fires back with Watson Analytics, which is a great visualization tool but also, crucially, is the expert. So no, they won’t let us skip the truly unique, potentially revolutionary features of a tool that analyzes the data automatically and can understand your questions.

Diving Into The Data with Watson Analytics Explore

Given that, let’s select “What is the contribution of Receptions over Week by Position?” Below you can see the results of a single click – a full featured visualization with a host of options for discovery and refinement. Let’s take a look at the interactive features:

The Explore Window explained.
The Explore Window explained.
  1. Let’s start with the big one first – the visualization. Clicking on visualization elements brings up an explanation of their contents and provides a menu for refinement of the data – custom calculations, filters, etc can be accessed by simply clicking on a bar, point, bubble or box. Awesome.
  2. Jumping to the top, here we find a number of other relationships Watson has uncovered automatically along with simplified visualizations representing the data. Clicking any of them will launch a full visualization like the one you see above.
  3. The question you asked is displayed here. Here the words in bold from earlier are interactive – want to swap “Rushing Yards” for “Receptions?” Simply click and select the data you want. The visualization automatically updates.
  4. The key is also interactive – clicking an element in the key opens a window to filter, calculate and otherwise manipulate that data set.
  5. Anytime you see this icon in Watson Analytics you know you have options. Here we can make more advanced changes to the visualization as a whole.
  6. You can create more than one visualization in an explore session – this is where they are managed. Create a new visualization, navigate through them, or display them all at once with these buttons.
  7. The pin icon works like you’d expect it to – it pins the visualization for sharing with other users or the Assemble dashboarding tool. Next to that is the “change visualization type” button for swapping to different visualizations using your current data and filters. Finally, the visualization display options allow you to change the style and color of this visualization.

From here we have complete freedom to explore the data – expert users can breathe a sigh of relief. IBM has taken a huge leap forward in terms of design – you interact with each element by clicking on it rather than through hidden menus.  Almost everything reacts as you expect it to. Explore offers a variety of tools to shape the data however for more serious data quality tasks you’ll want to use Assemble, which we’ll cover in a future post.

Watson Analytics Explore In Action

This is without a doubt the easiest query/visualization tool in the IBM Analytics stack. Cognos Workspace Advanced and Report Studio require significant technical skill and education to use effectively, as well as a data modeler to clean and present the data. Watson Analytics Explore lacks many of the enterprise features of Cognos but is staggeringly more advanced when it comes to usability, agility and high fidelity presentation. Behold!

So pretty, so fast.
So pretty, so fast.

 

What you see above took approximately seven minutes of work. I can make something that looks very similar – but not as sharp – in Cognos in approximately two hours. However, clicking any of these charts will drill down to a large, fully interactive version which is easily refined. That functionality in Cognos is more like a consulting engagement. To top it off, the bottom pane is acting as a global filter – I currently have the NFC North selected, but the entire data set can be instantly filtered with the results rendered in all visualizations…instantly!

IBM’s Analytical Agility

“Agile” is, at this late stage, a methodology, buzzword and parody bundled neatly together, and its insights are eye-rolling obvious but still necessary, especially in Business Intelligence. A good BI deployment leads to new questions and a great one leads to new questions that must be answered urgently. If after go live the users politely agree that all requirements have been met and the system is complete you must immediately stow the champagne because you have failed.

This is where IBM and Cognos most often fail, due more to outmoded thinking than technical limitations in the product. Cognos does not require a six month workflow to make a new graph but the vestigial processes of the waterfall era weigh it down. In a sort of perverse way Tableau’s greatest strength is that you simply cannot do that much. The temptation to drown it in requirements, governance and process is ignored.

This is where Watson Analytics Explore fills a huge gap for Big Blue. It is free and easy to use. It looks great. It interacts with the much more complex features of Watson Analytics but doesn’t require them and thus offers a very low barrier to entry. And it is agile in the way a great BI project is agile; it answers questions with questions and does so beautifully. Like Tableau, it’s design makes obvious how it is to be used: quickly, joyfully and unburdened.

 

Filed Under: Watson Analytics

Reader Interactions

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Copyright © 2021 · Atmosphere Pro on Genesis Framework · WordPress · Log in