Q&A: "Cool BI" List Includes Dashboards, Visual Discovery Tools
Popular analyst Cindi Howson discusses top BI innovations.
- By Linda L. Briggs
- January 24, 2012
In part 1 of our interview, well-known BI analyst and TDWI faculty member Cindi Howson discussed "Cool BI" innovations including mobile, search, and cloud/software-as-a-service (SaaS) technologies.
In the second and final part of our interview, Howson discusses more technologies on her annual "Cool BI" list of business intelligent innovations. Topping this year's list: advanced visualization and dashboards.
"Visual discovery and dashboards continue to be two of the hottest technologies," Howson says. "Hot because of their value and hot because of the confusion between them."
Howson is founder of the BI Scorecard Web site, which offers independent evaluations of leading BI products.
BI This Week: Do you view visual discovery tools as an important space in BI?
Cindi Howson: I think it's probably the most important space. Visual discovery and dashboards continue to be two of the hottest technologies I see -- hot because of their value and also because of the confusion between the two terms.
What's the difference between visual discovery tools and dashboards?
I look at visual discovery as the process of exploring your data and finding the hidden patterns. Dashboards are how you monitor things at a glance. You might present a visualization inside a dashboard, for example.
When I talk about these categories, I draw them as overlapping concentric circles, a Venn diagram, because you have visual discovery products such as Tableau, which can make a dashboard, but you also have visual discovery products such as SAS JMP, which cannot make dashboards. You also have dashboard vendors such as QlikView, whose tools can be used for exploration, but which I wouldn't explicitly position as a visual discovery tool.
What are some other examples of vendors doing interesting things in with visual discovery and dashboards?
TIBCO Spotfire version 4.0 just came out; they've added collaboration, as has QlikView in QlikTech version 11. Tableau is in a beta for version 7.0, so they are adding some enterprise capabilities, I would say. MicroStrategy just released Visual Insight. Cognos has been testing and doing demonstrations in the lab on a new visual discovery product that's not yet named. Microsoft has been demonstrating Crescent, a visual discovery tool expected to be out next year, and Oracle Exalytics has a new visual discovery tool and appliance that's in beta..
Some of these are platforms and some are specialized tools. Do you have advice for a company looking for a visual discovery tool? Is it better to go with an in-platform tool or a specialized tool?
I think the bottom line is that visual discovery and dashboards are two modules that should be part of everyone's BI tool portfolio. But within those two modules, you sometimes get better capabilities from specialty vendors than from the BI platform vendor. It depends on who the BI platform vendor is.
As of right now, for visual discovery, I'm seeing better products and capabilities in the specialty vendors, so you need to evaluate how important integration is versus better capabilities and faster time-to-value.
What about social media and BI? How well are companies addressing that market right now?
I think vendors are ahead of users here. There are two ways to think about it. There's the impact of collaboration and social networking on the BI tools, so BI tools are becoming more collaborative. That's one aspect. The other thing to consider is that it's a whole category of sentiment analysis and social media analysis, and that area is highly fragmented, with lots of specialty solutions.
Have the larger BI platform vendors integrated ways of working with social media?
Not yet. SAS has a product and Cognos has a product. SAP just announced a partnership with NetBase Beyond that, I can't think of any other big BI platform vendors that do.
Is that because users just don't know they need it, or they're not asking for it yet?
It's more that social media data is separate from the main BI environment and may only be viewed by the marketing department. Also, some people are still skeptical about the value of analyzing that data. Is it just comments from people who use Twitter to complain about your product? Most companies are struggling enough with the data sources that they already have.
The collaborative aspect within BI tools is more interesting because it's changing the products. In that case, I think the vendors are ahead of the customers there.
Big Data seems to be a popular phrase lately. Is that on your Cool BI list?
I'm still looking at it. I tend to focus on the business layer and on the business user interface aspects, and I'm not sure yet [where big data will fall]. Is it just another data source? What's the impact on the presentation layer?
Do you see vendors actively saying, "We need to address Big Data issues?"
I think there are two ways of viewing Big Data. Is Big Data synonymous just with Hadoop, or is it the case that there is more and more data to be analyzed -- how do we make sense of it all? I think Big Data can be viewed both ways, but a lot of people right now just think of it as synonymous with Hadoop.
To the extent that vendors are helping businesses analyze these new data sources and bigger data volumes, "Big Data" then is more than Hadoop integration. However, I am seeing new vendors pop up, such as Datameer, DataMarket, and Karmasphere.
Is predictive analytics on your Cool BI list this year?
It is, to the point that it integrates with the BI platform now. I don't look at predictive analytics as a separate category. To me, what's more interesting is analytics being integrated with the BI platform, whether that means pushing analytics into the database or making the results of those models more consumable in dashboards.
I don't think predictive analytics will ever be a mainstream technology that many people need to become experts in. I think it will continue to be a specialist task and set of tools. The focus is, how is it processed and how are the results consumed? I think that part can be mainstream.
Among vendors, who is doing interesting things with predictive analytics?
SAS released Rapid Predictive Modeler last year, which was simplifying the making of the models. They've been pushing some of the processing into the database, such as the Teradata Database. Then, the results of their models can be consumed inside a report or a dashboard. I think that's a good move.
There's Cognos, of course. With IBM's acquisition of SPSS, they've added some statistical functions directly within their business metadata layer, called the Framework Manager model. I think they're doing some interesting things. Again, Information Builders is doing some very interesting things based on open source R.