Monday, May 09, 2011

Visual Business Intelligence using MS BI

I'm reading: Visual Business Intelligence using MS BITweet this !
Business Intelligence has a lot of adjectives associated with it, and one that is generally not found in limelight is visual business intelligence. Those who are not aware of this term would even feel that it's just a hoax and no such category of BI exists, but many do not know that lack of attention to this form of BI is one of the major reasons why major BI projects fail. After this statement the question is, what is Visual BI and why I have never heard it in MS BI world ?

There are two groups of professionals:

1) Those who just believe in ideology of "Code is GOD". This group generally looks upon theorist as people who are dumb in knowledge as they do not deal with implementation day-in day-out.

2) Those who believe in associating implementation to methodical and design approach. Design approach is a very broad term and is not limited just to design patterns. User Experience, Information Representation, Visual Analytics does not just reside in theoretical whitepapers or books, but in practice they take a very high prominence when dealing with critical nature or volume of business as well as business users.

Cutting the theory and back to technology, visual business intelligence means how good you understand your data, purpose of data analysis, users who would be doing the analysis and based on this how you design the visual representation of this data on your reports and/or dashboards. Professionals who are seasoned with designing reports using SSRS / Crystal Reports / Business Objects / Actuate would call them reporting professionals, and provided they have a few years they would start considering them reporting analysts. In my views just as you use a reporting tool to report data from a cube, does not make you a report analyst. A report analyst role is a very key role in dashboard design and this is the role who would be mostly responsible for the Visual BI.

Visual BI is all about representing data in the best form suited to the use of the data, so that the users can make analysis using just the shape and eyes, without making any calculations to make the analysis. A report analyst would know whether to use a multivariate analysis using heatmap, or whether to use histogram for statistical analysis, or to use scatterplot for outliers analysis, or use a calculated series with line charts for trend analysis, etc. An ignorant report analyst might use 3D-charts to make the graphics appealing, but would overlook the factor that third dimension does not add any value in terms of data. If you carefully look at a 3D Bar Chart, there is a high probability that you would often get more confused while comparing bars as the slanted view does not provide enough clarity to compare size. This is where Visual BI comes to play, to make the graph useful to the extent that user would take all the required answers just by looking at the graph. For this right selection of graph for the right kind of analysis with the right kind of information representation is required.






SSRS 2008 R2 might not be as strong as it's competitors in terms of features, but with the new graphical and charting enhancements it has a lot to offer in terms of Visual BI. It depends on how informed you are on the visual design of the information and user experience aspect. SparkLines, DataBars, Interlacing effect, Pareto analysis based charts, Customizable gauges etc are highly effective tools, and provided you have the right knowledge to design information visualization using these tools, you can defeat a dashboard designed using some million dollar enterprise software suite. Visual BI is the success ingredient for an effective dashboard design, and in my opinion, one should focus and invest in learning reporting techniques than mere reporting tools. But if you fall in "Code is GOD" category, pray that the tool of your expertise survives forever !!

1 comment:

RightSideOfWrong said...

That's a brilliant post.Really an eye opener!

Related Posts with Thumbnails