Tim

Tim Kent's Blog

Win - Loss Graphs in Reporting Services 2008

As I mentioned previously, SSRS 2008 allows us to use some nice visualisation techniques in our reports.  Win\Loss graphs are used to show (wait for it......) the trend of wins and losses over time.  The usual example is the performance of a sports team during a season, but they are also useful for showing trends over time such as the performance of a KPI.  Here's another quick how-to on building this type of chart in SSRS 2008.

So first we need a dataset with three fields: Series, Time and Value.  The first two are fairly self-explanatory, for the third you need to transform your value so it appears as either a 1, -1 or 0 (win, lose, no result/tie).  The AdventureWorks built-in cube KPI's work well here so I'm going to shortcut slightly and use these.  If you're going to use a relational source you may need to do something a little cleverer.  For example:

SELECT
    [Month]
    ,[Country]
    ,CASE 
        WHEN [SalesAmount] > 10000
        THEN 1
        WHEN [SalesAmount] < 10000
        THEN -1
        WHEN [SalesAmount] IS NULL
        THEN 0
    END AS Winlose
FROM 
    [AdventureWorksDW].[dbo].[FactInternetSales]

We'll be using AdventureWorks (SSAS) "Internet Revenue" KPI.  Don't forget that when using an AS data source in SSRS you basically have to trick it into thinking it's relational and make sure you have your measures on Columns (as Chris Webb discusses here).  Here's our very simple MDX statement:

SELECT 
    KPITrend("Internet Revenue")
ON COLUMNS, 
    [Date].[Calendar].[Month].members 
    * 
    [Customer].[Customer Geography].[Country].members
ON ROWS FROM 
    [Adventure Works] 

Now lets add a chart - pick the standard column chart and format as follows:

  1. Right click and delete the chart title and legend;
  2. Right click on both the axis titles uncheck Show Axis Title
  3. Right click on the Y (vertical) axis and select Axis Properties.  Under Axis Options set Minimum to -1, Maximum to 1 and Cross At property to 0 (this makes sure our baseline runs between the "wins" and "losses").  Click OK;
  4. Right click again on the Y axis and uncheck Show Axis;
  5. Right click on the X axis and select Axis Properties.  Under Labels  check Hide Axis Labels and do the same for Major and Minor Tick Marks. Under Line set the width to 0.5 and the colour to light grey or similar.  Click OK.

You should now have something like this:

image

Now lets add some data:

  1. Drag your Time field from the Report Data pane into the category area;
  2. Drag your Value  field into the Data Field area;
  3. Right click on your series and select Series Properties.  Under Fill set the colour suitably.  In our example we're going to show wins as black and losses as red using a switch statement in an expression:
  4. =Switch
    (
        Fields!Internet_Revenue_Trend.Value = 1, "Black",
        Fields!Internet_Revenue_Trend.Value = -1, "Red"
    )

If you preview the chart now it should look something like this:

image

To finish off we need to add the chart inline to a table:

  1. Add a table and delete the third column;
  2. Set the DataSetName property to the same as your chart;
  3. Right click on the second row and add a group (using the adjacent above option).  Select your Series value (in our case: Country) from the drop down;
  4. Delete the bottom row;
  5. Drag your Series field into the First column, second cell;
  6. Drag your chart into the second column, second cell.

Finally add a bit of formatting to your table and you're done!

image

As usual the sample report can be downloaded here: WinLose.rdl (24.08 kb)

This also works in SSRS 2005 - though doesn't render quite a cleanly.

PerformancePoint Monitoring - Hide empty rows in a grid or chart by design

So as the dust settles and everyone calms down, it's time to move on and get back to what we're good at! Just in case anyone was worried, we'll continue to be at the cutting edge of PerformancePoint Services and the MS BI stack as well as looking at SharePoint in more detail.

I thought I'd post this one as it's come up with a couple of our customers recently and isn't entirely clear at first glance.  When you design an Analytic Chart or Grid in Dashboard Designer by default empty rows/columns (or series on a chart) are shown.  If you want them to be hidden when the user first opens the chart/grid then click the browse button at the bottom right of the design pane:

image

The preview window will appear.  Now click either of the two end icons on the toolbar (marked by the red rectangle below) to hide the empty rows/columns (or both)

image

When you click OK the setting will be saved into your design - you can verify this by going into the MDX view of the query and checking that "Non Empty" now appears in your MDX. 

You can use the same method to apply sorting by right-clicking on a column in the browse window and selecting the relevant sort function.

image

Of course if you are using an MDX query then use the NonEmpty/Non Empty functions directly.

As I've pointed out before, don't forget the built-in Dashboard Designer help which is pretty comprehensive

 

How to: Bullet Charts in Reporting Services 2008

Unlike SSRS 2005 where creating bullet graph takes a bit of jiggery-pokery, Reporting Services 2008 almost has them built-in (thanks the the acquisition of Dundas technology).  I say almost as it's actually using a linear gauge object which in the wrong hands allows you to put the likes of a thermometer on your reports to show just how "hot" your business is!  Luckily, there is a built-in bullet graph template for the gauge but it still requires a bit of customisation to make it look like a bullet graph as per Stephen Few's specifications.  However it's almost all formatting/setup.  Here follows a quick(-ish) how-to.

In most cases you'll be using this type of chart inline in a table, however it's easier to develop the graph object outside of the table then drop it in once you've got the formatting right.  To start with create yourself a datasource and build a dataset that has the following fields and one row for each bullet graph you want to display:

  • Category Name
  • Actual Value
  • Target Value
  • Bad Zone Max
  • Satisfactory Zone Max
  • Good Zone max
  • Optionally you may need a bad zone min if you have negative values

Next drop a gauge object onto your report and select the bullet graph template. 

image

You'll notice that the default is lots of jazzy shading and spangly borders.  We particularly need to get rid of these if we are using the graph inline as they will make the graph very hard to read when scaled down (and of course add no value in terms of visualising the key information anyway).  We'll also want to remove the scale if we are using inline - we can add it back into our column header later.

Right click on the graph and select gauge properties.  Select the Back fill section and set Gradient to solid and the colour to no colour.  Then select the Frame section and set the style to none.  Close the properties pane. Looking better already:

image

Next lets remove the scale - right click on the graph again, this time selecting Gauge and then Scale Properties.  Select the Labels section then check the Hide Scale Labels checkbox.  Then select the Major Checkmarks section and check the Hide Major checkmarks checkbox.  At this point also set the start and end margins to 1 in the Layout section and set the Maximum value of the scale to your good zone max field.  Close the properties panel again.  

Now lets deal with each component of the graph.  You'll notice that the orange bar is actually made up of two objects known as linear pointers.   Lets tidy them up and assign our Actual and Target values to those pointers.  First make sure you've set the relevant datasetname property for the graph.  Then right click on your graph again, gauge sub-menu and Pointer(LinearPointer1) properties.   Set the value to be your actual field from your dataset.  The go to the Pointer Fill section, set the fill style to solid and pick a suitable colour for your "bullet".  Set the Pointer width to 30 and finally select the shadow section and set the shadow offset to 0.  Do the same for your target pointer (LinearPointer2) but this time setting a width of 8 and an length of 80.

image

Now set the relevant properties for each of the zones.  This time select Range(LinearRange1) properties from your Gauge sub-menu.  Set the Start Range to 0 and the End range to your Bad Zone Max field.  Whilst you have the properties pane open, get rid of the range borders by setting the border style to None and the start and end widths to 60.  Repeat for range two but this time setting the Start Range to Bad Zone Max and the End Range to your Satisfactory Zone Max. The same for zone three but this time with Satisfactory Zone Max and Good Zone Max.

Now add a table to your report, set the datasetname to the same dataset as your graph.  Delete the third column and drop your Category field into the second row, first column and then add a grouping on the Category Field (right click on second row, group properties, add).  Then drop your chart into the second column, second cell.  You should have something that looks like this in design view:

image

Almost there.  One of the quirks of any inline chart is that it overrides the border of the cell it's placed in so you need to set the border for the GaugePanel to the same as your cell borders.  Finally lets add a scale to the header of column two.  Copy your chart from the cell and then paste it the header.  Remove all the components except the scale from the gauge panel by carefully highlighting and deleting.  Right click the Gauge in your header and select scale properties from the sub-menu.  Uncheck the Hide Scale Labels checkbox and the Hide Major checkmarks checkbox.  For the max value of the scale you need to make sure you use the max grouping for you good zone max rather than the default sum (=Max(Fields!Zone3.Value)).  In the layout section, set the position in gauge to 90, in the labels section set the placement to inside and distance from scale to 30.  Set the major tick mark placement to inside.  Phew!

Tidy up your fonts and general layout and you're good to go:

image

That's about it - if all that seems like a lot of work download the sample report and use that instead :)

HowTo2008Bullets.rdl (41.22 kb)

Double MVP award for Adatis

I was struggling to write this post without too much trumpet blowing, but luckily Matt has done that for us!

I'm very pleased to announce that both Sacha and myself have both been awarded Microsoft MVP for 2009.  Whilst this is somewhat surprising it's also nice to know that someone is taking notice of the extra effort that goes into blogging, presenting, answering forum questions, developing freeware like our PPS planning web part etc.

Thanks as ever to the outstanding Adatis team who've earned this just as much as we have.

What's New in PerformancePoint SP2?

Correction: My badly titled post suggested this was the list of fixes rather than new features.

Unless the read-mes (Planning and Monitoring) are still being updated, there's no real surprises in terms of new features in Service Pack 2 for PerformancePoint Server.  More about the fixes in the Knowledge Base (KB958291 KB960815)

Planning

Addition/Change

Description

Support for Windows Server 2008 Hyper-V™

You can now use PerformancePoint Server 2007 SP2 with Windows Server 2008 Hyper-V.

Hyper-V creates new opportunities for server virtualization. You can use Hyper-V to make more efficient use of system hardware and host operating system resources to reduce the overhead that is associated with virtualization.

For more information, see the PerformancePoint Server 2007 Hyper-V guide.

Monitoring

Addition/Change

Explanation/Description

You can now use SQL Server 2008 with PerformancePoint Server 2007 SP2.

Important: To use SQL Server 2008 with PerformancePoint Server, you must install PerformancePoint Server 2007 SP2 before you install SQL Server 2008.

You can now use the Show Details action on PerformancePoint reports that use data that is stored in SQL Server 2008 Analysis Services.

Show Details enables dashboard consumers to right-click in a cell or on a chart value and see the transaction-level details for that value.

PerformancePoint Server 2007 with SP2 now supports Windows Server 2008 Hyper-V.

Hyper-V creates new opportunities for server virtualization. You can use Hyper-V to make more efficient use of system hardware and host operating system resources to reduce the overhead associated with virtualization.

For more information, see the PerformancePoint Server 2007 Hyper-V guide (http://go.microsoft.com/?linkid=9639690).

You can now use Dashboard Designer on a computer that us running .NET Framework 3.5 alongside .NET Framework 2.0.

You must install .NET Framework 2.0 before you install .NET Framework 3.5.

You can now use PerformancePoint Server with domains that have apostrophes in their names

In previous versions of PerformancePoint Server, when a domain name included an apostrophe, the configuration tool failed for both Planning Server and Monitoring Server.

Scorecard key performance indicator (KPI) queries are improved.  

Timeout errors no longer occur with scorecard key performance indicators (KPIs) that use data that is stored in SQL Server 2005 Analysis Services.

Time Intelligence Post Formula filters now display the correct number of days for each month.

In previous versions of PerformancePoint Server, the calendar control for Time Intelligence Post Formula filters sometimes displayed 31 days for each month. This is no longer the case.

Time Intelligence filters now work on scorecard KPIs that use data that is stored in Analysis Services

In previous versions of PerformancePoint Server, some Time Intelligence expressions caused filters that were linked to KPIs to fail.

For example, when a compound expression such as (Day-7:Day-1) was used in a Time Intelligence Post Formula filter and that filter was linked to a KPI, an error message occurred. In PerformancePoint Server 2007 PS2, single and compound Time Intelligence expressions work with KPIs that use data that is stored in Analysis Services.

An Old New Blog

One of our colleagues in the UK BI world has recently started blogging again and is well worth subscribing to if you don't already.  About time too Mark!

Mark Hill - http://www.markhill.org/blog

When I get a spare moment (sometime in 2011 I think!) I'll update the Adatis blog roll with this and a few others.

 

 

We have moved (again!)

After a short spell in Woking, we have now moved to a more permanent location in Farnham, Surrey.   Please update your records:

Adatis Consulting Limited
A2 Endeavour Place
Coxbridge Business Park
Farnham
Surrey
GU10 5EH
United Kingdom

Maps can be found on our website

HowTo: Bullet Charts in Reporting Services 2005

In my last post, I showed a dashboard in SSRS 2005 that included sparklines, inline bar charts and bullet charts.  As I mentioned there is good MSDN article on how to do inline charts but the Bullet chart took a little more thought.

image

As one of my colleagues has pointed out, these aren't bullet charts in the truest sense that Stephen Few designed them as they don't show the "bullet's path" running through them, however I think they serve the same purpose albeit in a simplistic way

The charts are made using 100% stacked bars with three data values for each "zone" of your chart:

  • The section below your marker;
  • The marker itself (which of course only shows if it falls in that zone); and
  • The section above your marker.

So in the Speciality Bike Shop row in the example above, the three values that add up to the first zone of 25 are 11, 1 and 13 with the colour set as Dark Grey, Black and Dark Grey respectively.  Whereas as Value Added Reseller is made up of three zones with values 0, 0 and 25

Likewise, the 2nd zone for Speciality Bike Shop is made up of 0, 0 and 64 whereas VAR is 42, 1 and 22 this time with light grey, black, light grey

So how do we come with the numbers? It's just a case of calculating the size of the zones as a percentage of the whole and deciding whether the marker should show.  Lets follow the MDX query through to explain

First lets define our key measures: the zone limits, and the actual value for our marker.  I'm using constants for the zone sizes but you could easily have this coming from your data:

WITH 
//replace this constant with your Zone 1 measure    
MEMBER [Measures].[Zone1] AS
     300000

//replace this constant with your Upper Zone 2 measure
MEMBER [Measures].[Zone2] AS
     1067000

//replace this constant with your Upper Zone 3 measure
MEMBER [Measures].[Zone3] AS
     1200000
    
//Actual measure
MEMBER [Measures].[Actual] AS
    [Measures].[Reseller Sales Amount]

Next we need to calculate which zone the marker will fall in - we set a boolean for each which is also used as the marker for the chart.

//Are we in the zone 1
MEMBER [Measures].[ShowZone1Marker] AS
    IIF(
        [Measures].[Actual] < [Measures].[Zone1],
        1,
        0
    )
//Are we in Zone 2
MEMBER [Measures].[ShowZone2Marker] AS
    IIF(
        ([Measures].[Actual] >= [Measures].[Zone1]
        AND
        [Measures].[Actual] < [Measures].[Zone2]),
        1,
        0
    )

//Are we in Zone 3
MEMBER [Measures].[ShowZone3Marker] AS
    IIF(
        [Measures].[Actual] >= [Measures].[Zone2],
        1,
        0
    )

Then we need to calculate the size of the lower and upper parts of each zone, lower first: 

MEMBER [Measures].[Zone1a] AS
    (([Measures].[Actual]) * [Measures].[ShowZone1Marker]
    /
    [Measures].[Zone3]
    * 
    100)
    - 
    ([Measures].[ShowZone1Marker] * [Measures].[ShowZone1Marker])

the logic is:

The actual value subtract the previous zone limit (not required in zone 1),  multiplied by the boolean value of whether or not the marker falls in this zone as a percentage of the whole chart.  The bottom line removes one from the value to allow for the marker.

The upper part of zone 1:

MEMBER [Measures].[Zone1b] AS
    ([Measures].[Zone1] - ([Measures].[Actual] * [Measures].[ShowZone1Marker]))
    /
    [Measures].[Zone3]
    * 
    100

this time the logic is:

the zone size subtract the actual value multiplied by the boolean value of whether or not the marker falls in this zone as a percentage of the whole chart.

Repeat the same for subsequent zones and then bring it all together in a select statement on your cube

Just in case I've explained all this as badly as I think I have, you can download the sample project and have a look yourselves: BulletCharts.zip (7.42 kb)

And the scale - errr OK in case you hadn't already guessed, it's an image on the column header - hey what more do you want?the moon on a stick? ;) You can size the chart to fit the scale or create your own image.  The image is also included in the sample project.

All criticism/suggestion for improvement welcome.

Bullet Charts and Sparklines in Reporting Services 2005

Here at Adatis we're big fans of Reporting Services.  Yep, it has it's quirks and frustrations (don't talk to me about excel export!) but doesn't deserve all the criticism it gets in our opinion.  You can create some great looking reports with not a lot of extra effort over the defaults (though how many times have you seen a "slate" style report called "Report1" go into production!)

We're also fans of what some might call the more "trendy" visualisation techniques such as sparklines and bullet charts and with Microsoft acquiring Dundas technology for use in SQL 2008, SSRS has some much improved functionality in this area.  I'll be taking a look at some of these new features in a future post.

Whilst SSRS 2005 does have it's limitations, with a little effort you can still create some very effective information visualisations.  The dashboard below is built using completely standard functionality

image

Adding sparklines and in-table bar charts is very simple (and pretty clever IMO).  This MSDN Article covers how to do it in detail but in essence you simply add a line or bar chart into a cell in your data table.

Whilst the bullet chart (in the Sales Channel Performance region of the dashboard) may be simplistic compared to those available in SSRS 2008 or MicroCharts from XLCubed (from whom I nicked the styling for our dashboard) it still provides a very clear visualisation of the data.  This required a little more trickery in the underlying query but is also just a chart in a table cell - this time a 100% stacked bar.  I've explained how this works here