Sunday, January 27, 2013

[OBIEE11g] - Choosing the Right OBIEE Visualization


Presenting business insights includes choosing the appropriate data visualizations to tell a story or make a key point. That may seem easy but choosing the best one is quite powerful. Getting the best use of data visualizations is and often overlooked but it is important part of bi design. Let’s looks at some of the best practices with OBIEE and present some common mistakes to keep in mind.

Using color

Color is an import part of any BI tool. Developers should consciously choose a color palate for most reports and dashboards. Color must be consider when conditional formatting, creating map views or graphs, and for the overall feel of the dashboard. Colors schemes fall into 3 categories.



Which Chart Should I chose?

Bar Charts

Bar charts should always start with zero and show nominal data values in comparison toeach other. They can be vertical or horizontal. Design should always try to avoid horizontal scrolling which could dictate the orientation of the bar chart. Utilize features like section scrolling or graph prompts to maximize dashboard real estate and offer the cleanest looking charts. The bar chart to the right starts at zero, compares products, allows the user to section slide for month over month numbers, and allows for prompting on the company using a graph prompt.

Stacked Bar Chart

Stack bar graphs can be confusing if not used appropriately. Stacking numbers like percentages, or loosely related dimensions can lead to misleading results. The total is the most clearly identified number of the display and should be the most relevant fact on display. It is best practice to set the largest stack on the bottom as much as possible. Colors or patterns should be easily distinguished and use a qualitative scheme. Area charts show the stacked relationships (totals) best flowing over time.

Pie Charts

Most often, pie charts are misused to communicate part-to-whole scenarios where line or bar charts would be much more effective. They should not be done in #D, have a limited number of slices, and be used to show percentage of the whole. Many visualization experts dislike them as they tend to be misused.


Scatter Plots and Bubble Graphs

Scatter plots are great options for displaying relationships between two quantitative variables, even with exceptionally large sets of data. Best practices around scatter plots include removing fill color where possible, visually identifying groups when multiple groups are plotted together (shapes, images, shades of color), displaying trend lines and using trellis charts to reduce complexity. Bubble charts limit the number of points that can be plotted but allow for a 3rd metric to be compared on the same chart.


Trellis Charts

Trellis charts are a small series of charts that much like sparklines also provide a very fast visual comparison of trends over time periods. In 11.1.1.6.4 they are available as simple or advanced trellis charts. Think of a pivot table on steroids, where you can show charts in context of a 2 axis pivot table. It really allows for the maximization of data consumption on one page. In the example below the columns represent time periods during the day, the rows represent the flights distance in 3 buckets, and the bubble chart shows 4 different metrics in each in cell. The color represents performance rating, the vertical axis shows the number of routes, the horizontal axis shows the % late, and the bubble size is the number of flights. It paints a picture of shorts routes having more flights and diminishing performance as the day progresses 
These are a few examples of visualizations in OBIEE. Choosing the one that best tells the storey is the key to good dashboard design.