Information delivery leading practices - IBM Cognos

The IBM Cognos Business Intelligence (BI) product provides a unified, interactive workspace for business users to create their view on data by combining all types of information and to personalize content to provide unique insights and to deliver faster business decisions.

With IBM Cognos Business Insight, business users can use flexible dashboards and reports. Business users can customize existing dashboards and change them in a way that answers their question or they can build completely new dashboards. From the user perspective, a dashboard needs to be an uncluttered display of relevant data, that provides an at-a-glace view of business performance with a simple and visually compelling way of presenting data.

To achieve this goal, consider the following recommended approaches:

• Use reports that focus on data that is of interest to the user.

The IBM Cognos Business Insight workspace is designed to allow a user to focus on an area for analysis, unlike production reporting, which is typically static in nature. Therefore, reports that are added to a dashboard are focused and summarized. You can include any report in IBM Cognos Business Insight. However, detailed reports with which the user does need to interact should not be included in an IBM Cognos Business Insight workspace.

• Use saved output or views.

With IBM Cognos Business Insight, you can use saved output as part of the workspace. If saved output is included in a dashboard, users can change the version that they want to see, set the dashboard to open with the latest version, or execute the report.

• Use atomic-level reports or purpose built parts.

When creating a new report for a dashboard, build reports at an atomic level, so that a report contains one object, such as a list or chart. Turn off the headers and footers, because they can confuse the consumer. Keep reports and report parts that are candidates to be included in IBM Cognos Business Insight workspace in a specific series of folders so that they can be found easily.

• Change the default names of parts or components.

Each of the parts or components in a report has a default name. For Lists, the default name is typically List1 or, if two lists are in the report List1 and List11. A good practice is to change the default names to a name that is more meaningful to users. Prompts do not have a default name and, as such, do not display in a Content unless you give them a nam

• Do not overwhelm users with charts.

When using a chart to represent the data, display only the data that is relevant for the users. For example, avoid multi-color backgrounds or third-dimensions added to bars or lines unless these features provide meaningful information.

• Reuse prompts.

You can use one prompt to apply the same filter to more than one report on the workspace. For example, if a report contains a date prompt, you can use that prompt to apply a filter to that report only. Alternatively, the other reports that have dates, for example the year, can respond to a year selected from the first report, as long as they share the same caption or dimension.

• Use My Favorites.

For quick access to the dashboards that you use on a regular basis, add them to My Favorites. After you save them, you can open the favorite dashboard or report from the Getting Started Page or the Content tab.

• Use workarounds for the dashboard printing.

You cannot print a dashboard directly from IBM Cognos Business Insight. However, you can save a report widget to a PDF file and then print the PDF file, or you can press Ctrl+P to use the web browser to print the information that is displayed on the screen.

• Use multiple dashboards.

If you want to see multiple dashboards at the same time, use the multi-tab option of the web browser. For that purpose, you need to open IBM Cognos Business Insight in chrome mode (with the web browser showing toolbars and menus).

• Avoid overlapping hidden widgets, and use appropriately-sized widgets.

Widgets can get “hidden” under other widgets, so try not to overlap them. Also, do not make a widget bigger than it needs to be. Use the “Fit all Widgets to Window” option after you add all the widgets to the dashboard.

• Use mobile support.

When using a web page in a workspace, try using the URL and adding /m to the URL to get the mobile version This option typically renders smaller results that fit better in a workspace. You can create several different types of reports and then use them in the dashboard.

List reports

A list report is a type of report that displays detailed information, such as product orders or a customer list. simple list report.

Simple list report

A list report shows data in rows and columns, where each column shows all the values for that item in the database. You can also group data in a list report by one or more columns, add summaries, or include headers or footers to provide additional information.

Crosstabs

Like list reports, crosstab reports (also called matrix reports) are reports that show data in rows and columns. However, the values at the intersection points of rows and columns show summarized information rather than detailed information. crosstab report.

Crosstab Report

You can create a nested crosstab by adding more than one item to rows or columns, or you can report on more than one measure. You can add any data that can be aggregated to the body of the crosstab as a measure. Measures define that data is reported, such as revenue, quantity, return quantity.

Charts

Charts are the visual representation of quantitative information. They reveal trends and relationships between values that are not evident in lists or crosstab reports. For example, you can create a report that visually compares actual sales and planned sales or that shows the percentage share of product lines in the total revenue of the company.

IBM Cognos BI includes the following chart types:

• Column charts
• Line charts
• Pie charts
• Bar charts
• Area charts
• Point charts
• Combination charts
• Scatter charts
• Bubble charts
• Bullet charts
• Gauge charts
• Pareto charts
• Progressive charts
• Quadrant charts
• Marimekko charts
• Radar charts
• Win-loss charts
• Polar charts

Column and bar charts

These charts show trends over time or compare discrete data. Column charts present data using vertical objects, and bar charts use horizontal objects. These charts make it easy to compare individual values just by comparing the heights or lengths of two bars. This type of chart is useful when you want to see how individual values are ranked from highest to lowest level where sorted column or bar charts are best.

Simple column chart

You can use more complex bar or column charts to display part-to-whole relationships as a stacked bar chart.

Stacked bar chart

Line, area, and point charts

Line charts are similar to column charts, but instead of using columns they plot data at regular points and connect them by lines. If you are interested only in a trend line but not individual values or when you are comparing many data series, this chart is a good choice. For example, Figure shows the distribution of gross profit over all the months in a year for different order methods.

Line chart

You can accomplish the same result with the area chart where, instead of having lines, areas below the lines are filled with colors or patterns.

Area chart

As for column and bar charts, you can use complex area charts (stacked area charts) to show the relationship of parts to the whole. Typically, it is best not to use stacked line charts because they are difficult to distinguish from unstacked line charts when there are multiple data series.

Another variation of a line chart is a point chart. A point chart is similar to a line chart, but the points on the chart are not connected with lines. Just the data points are shown.

Point chart

Combination charts

Combination charts are a combination of the charts mentioned previously. They plot multiple data series using columns, areas, or lines all within one chart. They are useful for highlighting relationships between the various data series. Figure shows a combination chart that displays revenue and gross profit for marketing campaigns.

Combination chart

Scatter, bubble, and quadrant charts

Scatter and bubble charts plot two measures along the same scale. Their purpose is to show correlations between two sets of data (measures). For example, Figure shows the relationship between quantity sold and return quantity for each product line.

Scatter chart

Using this chart makes it easy to discern the following patterns:

• Linear trends: Either positive trends (points are going up from left to rightin a pattern that looks like a line) that indicate a positive correlation between two measures or a negative trend (points are going up from left to right)
• Non-linear trends: Points in a pattern to form a curved line, indicating positive or negative correlation
• Randomness: Points arranged randomly, indicating that there is no correlation between two measures
• Concentrations: Points appear in particular area of the chart, for example in upper-left corner, which indicates many product or product lines with a small number of items sold and a high number of returns
• Exceptions: Points stand out from the remaining pattern, indicating anomalies, as in the previous chart example In this chart, the Outdoor Protection product line has a higher number of returns than other product lines.

Bubble charts are similar to scatter charts but contain one additional piece of information the size of the bubble represents the third measure. These charts are used usually for showing financial data. Note that bubble charts are not supported for Microsoft Excel output.

You can create more complex bubble (or scatter) charts by adding a forth measure to the chart by specifying that the data point appears in different colors based on that measure. Figure shows an example of a chart with a correlation between Unit Sale Price and Unit Cost. The size of the bubbles shows the Gross Profit and the color of the bubbles shows whether the quantity is less than 1,000,000 (red) between 1,000,000 and 20,000,000 (yellow) or greater than 20,000,000 (green).

Bubble chart

Quadrant charts are in fact bubble charts with a background that is divided into four equal sections. Legacy quadrant charts use baselines to create quadrants, and current default charts use colored regions. You can change the size of the quadrants. You can use quadrant charts to present data that can be categorized into quadrants, such as strengths, weaknesses, opportunities, and threats (a SWOT) analysis.

Quadrant chart

Pie and donut charts

Pie and donut charts are used to show the relationship of parts to the whole by using segments of a circle. To show actual values, a stacked bar chart or a column chart provides a better option. Also, pie charts are not a good choice for a chart if you have measures that have zero or negative values. Figure shows a pie chart showing proportions of advertising costs.

Pie chart

Note that reports in PDF format or HTML format show a maximum of 16 pie charts.

Bullet charts

Bullet charts are one variation of a bar charts. Bullet charts shows a primary measure, in comparison to one or more other measures. It also relates the compared measures against colored regions in the background that provide additional qualitative measurements, such as good, satisfactory, and poor.

Bullet chart

Because they deliver compact information and do not need too much space on a dashboard, you can add bullet charts to other report objects, such as list reports.

Combination of a bullet chart and a list report

Gauge charts

Gauge charts (also known as dial charts or speedometer charts) are similar to bullet charts in that they also compare multiple measures but use needles to show values. Reading a value from a gauge chart is as easy as reading a value on a dial, and each value is compared to a colored data range.

Gauge charts are a better option than a bullet chart in the case where you need to compare more than two values (measures).

These charts are usually used to show the KPIs in executive dashboards. Note that PDF output and HTML output of reports are limited to show up to 16 gauge charts. These charts are not available for Microsoft Excel output.

Figure shows how to compare three measures (product cost, planned revenue, and revenue) on the same gauge chart.

Gauge chart

Pareto charts

Pareto charts rank categories from the most frequent to the least frequent. They include a cumulation line that shows the percentage of the accumulated total of all the columns or bars. You can use these charts for quality control data, so that you can identify and reduce the primary cause of problems. You can create before and after comparisons of Pareto charts to show the impact of corrective actions. Figure shows an example of a Pareto chart showing the gross profit (in millions) for regions by product lines.

Pareto chart

Progressive column charts

Progressive charts (or waterfall charts) are a variation of column or stacked charts, with each segment of a single tack displaced vertically from the next segment. Progressive charts, as well as stacked bar or column charts and pie charts, are useful for emphasizing the contribution of the individual segments to the whole. These charts are not supported for Microsoft Excel output.

Figure shows the contribution of each Product Line to Gross Profit (in millions).

Progressive chart

Marimekko charts

Marimekko charts are stacked charts in which the width of a column is proportional to the total of the column’s values. Individual segment height is a percentage of the respective column total value. Figure shows the contribution of return quantity of returned items for product lines by order methods.

Marimekko chart

Radar charts

Radar charts compare several values along multiple axis that start at the center of the chart forming a radial figure. These charts are useful if you want to compare a couple of variations against the same set of variables or to compare multiple measures. This type of chart is also useful for spotting anomalies or outliers. Figure shows an example of a radar chart that compares revenue by product lines for different retailers.

Radar Charts

Win-loss charts

Win-loss charts are microcharts that use the following measures:

• The default measure
• The win-loss measure

The win-loss measure is the measure or calculation that you define. You can also define the default measure, if necessary. You can use these charts for visualizing the win-loss trends, for example the months that have revenue over a certain threshold.

Win-loss chart

Polar charts

Polar charts are circular charts that use values and angles to show information as polar coordinates.

Figure shows the revenue and quantity for each product line. The distance along the radial axis represents quantity, and the angle around the polar axis represents revenue.

Polar chart

Baselines and trendlines

Baselines or trendlines provide additional details on a chart. Baselines are horizontal or vertical lines that cut through a chart to indicate major divisions in the data. For example, you can add a baseline to show a sales target or break-even point. Each baseline represents a value on an axis.

Example of baseline added to a chart

Baseline options (depending on the type of chart where baselines are available) include:

• Numeric value: Static numeric value
• Mean: Statistical mean plus or minus a number of standard deviations based on all charted data values on the specified axis
• Percentile (%): Specified percentile
• Percent on Axis (%): Percentage of the full range of the axis

Trendlines graphically illustrate trends in data series and are commonly used when charting predictions. A trendline is typically a line or curve that connects or passes through two or more points in the series, displaying a trend. You can add trendlines to bar, line, area, bubble, and scatter charts.

Figure shows an example of adding a polynomial trendline to a chart displaying revenue by product lines over time to see the trend.

Example of a trendline added to a chart

The following trendlines are available:

• Linear, for data values that increase or decrease along a straight line at a constant rate (for example revenue that increases over time period)
• Polynomial, for data values that both increase and decrease
• Logarithm, for data values that increase or decrease rapidly and then level out
• Moving average, for data values that fluctuate and you want to smooth out the exceptions to see trends