Let’s begin with the Bubble Chart method. Recall that the Bubble Chart is identical to the XY Chart, except it allows for a third variable to define the size of each point. When I first began investigating Venn diagrams in Xcelsius, the Bubble Chart seemed like a natural starting point since it already does half the work for you. That is, it automatically resizes the “bubbles” to reflect the proportional difference between the two groups. The only thing I had to do was build Excel formulas to shift the bubbles together in order to show the correct amount of overlap. Unfortunately, that second part is very complex and makes for a big disadvantage compared to the Google Charts API.

It turns out the Bubble Chart object is not exactly consistent or easy to predict. Its movement and behavior varies depending on the overall size of the object and the size of the individual bubbles. Therefore, it takes some work to force the chart to display the correct overlap. What I did was arrange a series of experiments where I changed the parameters of the Venn diagram and then measured how much shift I had to apply in order to achieve the right overlap. I then took that data and used it to build a regression model. Then I plugged the model equation into my Excel spreadsheet and used it to calculate the amount of horizontal shift I needed in order to see the correct “bubble” overlap. Sure, it works, but it’s a lot more effort compared to calling up a simple Google Chart URL. Also, this only produces a Venn diagram with 2 sets, while Google Charts can easily produce a 3-set Venn diagram.

However, the Bubble Chart approach does have some advantages over Google Charts. Perhaps the most significant advantage is that Bubble Charts do not require an internet connection and can be used in a fully offline mode. When using Google Charts, your user must be connected to the internet and have the ability to access http://chart.apis.google.com

There’s also a potential security advantage to using Bubble Charts. If your organization has strict security protocols, it may not be permissible for you to pass information to an external party (i.e. Google). However, you might be able to get around that by opting not to pass data labels through the Google Charts URL and instead handling the labels from within Xcelsius. This means that you would only be passing simple aggregate numbers to Google without a corresponding label to indicate what that data is or where it came from.

In my opinion, you’ll always want to lean towards using the Google Charts API. It’s more flexible, easier to use, and it has more features than you could produce using a hacked Bubble Chart. However, if you need your dashboard to run offline or you have a security rule which prevents the transmission of any data whatsoever to a 3^rd party, then you’ll need to use a Bubble Chart.

It’s hard to imagine producing a better display of that information using any other charting method. In other words, if you’re creating a visualization to compare overlapping groups, the Venn diagram option is probably at the top of your list. What makes a Venn diagram so powerful is its simplicity. It’s very easy for both technical and non-technical groups to view the chart and interpret the information. Yet, in spite of all this, Venn diagrams are almost always absent in reporting tools. Certainly this is true for the three major Business Objects reporting tools – Crystal Reports, Web Intelligence, and Xcelsius – which have never offered an out-of-the-box Venn diagram charting feature.

I recently became interested in building Venn diagrams using Xcelsius. I knew there would be significant value in providing this type of visualization to report users, especially executive level users. I also felt it would create a new analytical perspective since Venn diagrams show the type of metrics that are not often presented well in business reports. I thought there had to be some trick or workaround I could use to get the job done, so I began tinkering.

The first thing I learned along the way is that Venn diagrams are not as easy to build as you might think. They look simple on the surface, but there’s actually some fairly complex geometry going on behind the scenes. (Do a Google search for “circle-circle intersections” to see what I’m talking about.)

After some testing, I developed an initial solution using the Bubble Chart component in Xcelsius 2008. The Bubble Chart is identical to the XY Chart, except there’s a third variable that defines the size of each point. Look again at my simple example above. You’ll see that the chart basically did two things. First, it sized the circles to reflect the proportional difference between the two groups and then it moved the circles together to reflect the overlap. Well, the Bubble Chart already does that first part by itself. So my solution used the existing Bubble Chart functionality to automatically determine the correct dimensions for each circle. Then, I set up Excel formulas to dynamically calculate the X value (horizontal position) for each circle so that the correct amount of overlap was displayed.

Overall, I was happy with this solution, but there were some drawbacks.   It was a bit clunky due to the math required to control the movement of the  Bubble Chart and to calculate the overlap. Also, as you can see from the  screenshot above, Xcelsius has a limit on how big it allows the individual “bubbles” to grow in a Bubble Chart. So the end result is a very small SWF file that you would need to import into another document (e.g. another Xcelsius dashboard or a PowerPoint presentation) and then re-size from there.

Shortly after that initial experiment, I set about improving the process.  That’s when I stumbled across the Google Chart API. The Google Chart API is a simple, easy-to-use web-based chart creation tool provided free by Google. The API allows you to pass a few parameters via an HTTP request which prompts Google to return a PNG image containing your requested chart. Several chart types are supported, including Venn diagrams.   Check out the Google Chart Wizard for more information on the supported charts and formatting options.

Remember the simple Venn diagram example I showed at the beginning? That was actually created using the Google Chart API! The URL string below shows the HTTP request that was used to produce the chart. To see how it works, copy and paste the URL into your browser. You should get the same chart that was shown above. Just for practice, try to adjust the parameters defined in the URL string, refresh the page, and watch how the chart changes.

• chs=600×400: Dimensions of the PNG image file.

 

• cht=v: This tells the Google Chart API that you want to create a Venn diagram. Other chart types include Line Charts (cht = lc), Pie Charts (cht = p), and so on.

• chds=0,5000: he minimum and maximum values allowable for charting. Keep the minimum at zero. Depending on your data, you may need to increase the Maximum value.

• chd=t:66,100,-1,33,-1,-1,-1: These numbers specify the size and overlaps for all groups. The ordering is as follows: Group 1 size, Group 2 size, Group 3 size, Group 1 & 2 overlap amount, Group 1 & 3 overlap amount, Group 2 & 3 overlap amount, Group 1 & 2 & 3 overlap amount. In this case, we only had two groups to display, so all of the parameter slots reserved for the third group were set to “-1”. If you have a third group, then those “-1” values would be replaced with actual amounts. The Google Chart API provides for a maximum of 3 groups in Venn diagrams.

• chdl=Customers+who+purchased+Product+A|Customers+who+purchased+Product+B: This defines the group labels that will be displayed in the legend. If you have a third group, you’ll add another pipe delimeter at the end, followed by the desired group label.

• chdlp=b: This tells the Google Chart API to place the legend at the bottom of the chart. “b” is the code for “bottom”. The other codes are “l” for left and “t” for top. If you exclude this parameter from the URL string, the legend will be displayed on the right by default.

In addition to those parameters, there are even more you can use to further customize the look and feel of your Venn diagram. You can adjust the margins, change the colors, add a title, and so on. If you’re interested, the Google Chart API links above describe all the options.

So how can you display Venn diagrams in Xcelsius using the Google Chart API? Actually, it’s very simple. You’ll need your Excel data model populated with numbers showing the size and overlap for all groups you wish to chart.  Then create string concatenation formulas in Excel to build the URL string piece by piece. Finally, add a Slide Show object to the canvas and bind it to the cell containing your URL string. And that’s it. You should now have a fully functional, dynamic Venn diagram in Xcelsius.

The dashboard below is a sample I created to demonstrate this functionality. It allows you to adjust the group sizes, overlaps, and labels.  As you adjust the values, notice that the chart changes dynamically just as you would expect from any other Xcelsius component. Also, the dashboard displays two groups by default; however, you can click the checkbox on the left side to incorporate a third group. Also, keep in mind you cannot use this approach in a completely offline format. Your dashboard user will need a connection to the internet and the ability to call the Google Chart URL.

Here is the working dashboard.  You can change the values and see the chart change.

There are certainly pros and cons to both solutions – the Google Chart API and my original Bubble Chart solution – and I hope to blog about those in the future. However, I think the Google Chart API provides a better overall solution. It’s clean and simple and it provides the functionality to incorporate a third group.

Finally, I encourage you to experiment with Venn diagrams in your own Xcelsius dashboards. Play around with the options to see what works best for you. Feel free to share your experiences and feedback in the comments.

Although SAP has been very heavily promoting the Explorer product for the last couple of years, the adoption of this newer search based query tool has been slow. There are many reasons for this but the one that stands out the most is, that as a Business Intelligence tool, it is limited.  While SAP BusinessObjects offers a number of Business Intelligence tools for different functions, the majority of customers prefer to pick just one or two for mass user consumption and they will tend to pick tools that can provide the widest range of functionality.

I recently conducted an ad-hoc survey of about 200 BusinessObjects customers in the U.S. encompassing small, medium and large companies and found that the top BI tools in use were Web Intelligence, Crystal Reports, Desktop Intelligence and Xcelsius. When looking at which tools they had most recently added in the last 3 years and were strategic for their future BI initiatives, it was Web Intelligence and Xcelsius.

Web Intelligence is seen as an end-to-end BI reporting and query analysis tool. It has both a web based version and a rich client version to be used off-line. It has good integration with Microsoft through Live Office and use of web parts and overall it has a rich array of functionality for self-service BI.

Xcelsius is probably the best BI data visualization and dashboard tool on the market today. It allows customized development of complete end-to-end BI solutions. It integrates with Microsoft, Adobe, Email, Google, Web as well as a host of other BI solutions. It also has a rich offering of add-on components and functionality to enable write back, alerts and mobile integration.

When we look at SAP Explorer, it is not really in the same category as Web Intelligence and Xcelsius. Explorer is a search web-based query tool for simple end-user ad-hoc analysis with limited reporting and customization ability. It can integrate with other BusinessObjects tools like Web Intelligence and Xcelsius and allows export to Excel but there is no real integration with other applications. It is a great solution for “one off” BI applications but not for end-to-end solutions.

SAP has been previewing the next potential edition of Explorer known as Exploration Views since the beginning of the year and this allows more end-user control and customization of the views.

Both Explorer and Exploration Views demonstrate very nicely and it is easy to get excited when seeing them but then the same could be said for Web Intelligence and Xcelsius.

I am sure we will continue to see more of Explorer but I see the future strength of the SAP BusinessObjects BI solutions remaining with Web Intelligence and Xcelsius for the foreseeable future.

Image Credit: http://www.presentermedia.com

This year’s “Best Business Dashboard” award went to Rural Metro Corporation for their Daily Business Insight dashboard which was developed to show daily insight into operational, billing and financial data from different systems delivered in a simple and effective way. The interface was designed in a format that made sense to management and provided the ability to coordinate the distribution of matching Web Intelligence reports to their mobile devices. One of the key benefits of the dashboard was that it helped to create a consistent means of communication across the business.

The “Most Innovative Dashboard” award was won by McKesson Specialty Care Solutions for a dashboard they developed to provide a single view of all program enrollment and fill activity for a key client to quickly access the data they need in order to respond to internal data requests, in a timely, efficient and accurate manner. The dashboard made great use of different Xcelsius components including color coded maps, selection tables and a variety of different charts while using tabs and selection buttons to make the navigation both simple and powerful. The dashboard resulted in timely and accurate turnaround to requests that benefitted both McKesson and their client.

The “Most Valuable Dashboard” award was claimed by Banner Health for their automated dashboard used to analyze and monitor patient throughput within designated Emergency Departments within the Banner Health system. The dashboard is able to access a large number of key operational and clinical metrics providing standardization and trust in the data throughout the organization. One of the key benefits was an improvement in overall patient satisfaction by reducing the wait time between a patient’s initial arrival time and examination by a physician.

The fourth award this year was renamed the “Most Inspired Dashboard” award and was won by Sohar Aluminum from Oman for their dashboard to bring about accountability, transparency in daily work and ultimately consistent, accurate data. The dashboard was inspirational in both its design and functionality by making use of write back functionality enabling supervisors to enter their actual and target results through the dashboard itself. In addition, other alert dashboards were created that are automatically scheduled and emailed to supervisors who are late entering their data. The dashboard has been so successful that it has actually brought about a culture change in the way people look at data and processes at the company.

After the awards, many companies present vowed to enter even more amazing dashboards for the 2012 awards and I can’t wait to see them.

As I walked around the showcase showing Xcelsius dashboards running on an iPad, people stopped with looks of disbelief mixed with wonder.

“How are you doing that?”

“Xcelsius doesn’t run on an iPad or an iPhone”

Well, I am happy to report that with MyBI Mobile, not only does Xcelsius run on an iPad and an iPhone, but also on an Android. And it not only runs but it sings too! It has a fabulous simple menu system to serve up different content on the device and that content can also be Crystal Reports, Web Intelligence, Desktop Intelligence in addition to Xcelsius. With Web Intelligence you can also create new reports from the mobile device – that’s cool! When viewing and interacting with the content, you can perform a “pinch and zoom” action on any part of the screen to automatically enlarge that area of the dashboard or report.

While many companies have invested in building both dazzling and effective Xcelsius dashboard business intelligence solutions, they have often been challenged when trying to find ways to move these to mobile devices. The operating system for the IPhone and IPad (iOS) does not support Flash and SWF outputs natively but with MyBI mobile the barriers have been lifted. You can now take your key Xcelsius dashboards and just serve them up on iPads, iPhones and Androids without changing them and interact with all the features just the same as on the PC.

I cannot express enough how incredible this is and what it means to the millions of Xcelsius users around the world who want to take their dashboards mobile  - you are free of your shackles and you can now deliver Xcelsius to the most popular mobile smart devices out there.

I recently heard there are an estimated 200 million iOS devices in use today and now all those users can enjoy the full functionality of the amazing data visualization of Xcelsius through MyBI Mobile.

If the excitement at IBIS is anything to go by, you will certainly be hearing and seeing a lot more of this solution in the coming months so stay tuned.

Question:  Working with a data set consisting of 13months of data with one row for each day (so max 380 rows) and needing to show only monthly detail/summary data in the dashboard, which of the following would you consider to be the better approach to yield the best performance?

1.  Pull 380 rows in cache with date in one column -> pass STARTDATE and ENDDATE from the dashboard and pull every month’s data by writing a cache query.   When the user selects a month for display in the dashboard, Excel will calculate the start and end date of the month and pass those to the cache query with a WHERE clause specifying BETWEEN ‘@STARTDATE’ and ‘@ENDDATE’

Or,

2.  Load 380 rows in the dashboard Excel backend -> calculate the month from the date column within Excel then use a hidden filter to filter the month’s row based on the calculated month in the previous step and calculate the sum and average within excel?

Answer:  As a general rule, if the data set is 200 rows or less, I would say to go with Option 2.  If it is more than 200 rows, I would say you will get better performance with the cache query.  Of course, this also depends on the cumulative total # of rows being loaded into the dashboard.  For example, if you are planning on adding 10+ additional 200 row data sets, then I would definitely go with Option #1.

With too much data in the backend, Xcelsius performance suffers.    InfoBurst allows us to cache a large set of data, and then use a Cache Query to extract just the portion we need to view on the dashboard, as the user makes selections.  This is a tremendous help to the developer who wants to boost Xcelsius dashboard performance.

My most recent opportunity was spending a month in the exciting country of Oman, which, despite the recent unrest in the Middle East, proved to be a welcoming, progressive and abundantly interesting country.  Apart from the opportunity to explore this part of the world, it also proved to be an exciting work project, as the dashboards I developed for my customer there included several innovative solutions. 

Upon arrival, I found that the customer had 52 KPIs that were being tracked weekly & monthly in over 60 different Excel spreadsheets, maintained by dozens of people in various locations.  The solution we put together for the customer allowed them to begin maintaining all the KPIs in a database (as it should be), by using InfoBurst to write back values to a database table.  Using Xcelsius as a data entry tool, we created two different dashboards for entering metric values which, with the push of a button, ran update queries back to our database table, which immediately refreshed and presented the data in the main dashboard for the General Managers to view.  This solution provided a new precedence for accountability and consistency across the board.

 There were two additional innovations to this solution which I am excited about.  Both of these involved automatically emailing a mini dashboard as an Event based Alert.  The first mini dashboard was emailed to the project team if a user modified one of the KPI targets.  Key users were allowed to change a target, but the team wanted to be aware when this happened.  Using InfoBurst, we created a query based event that checked every 30 minutes for a condition to be true (a target change).  Whenever that condition is true, InfoBurst initiates a burst of a mini dashboard (.swf file) to the project team, with the following information:

·         Which KPI target was changed

·         What is the new KPI target value

·         Who changed the target

·         What time did this change occur

The second innovative Alert notification mini dashboard involved the time sensitive nature of getting all the KPIs entered by given dates.  Schedules were created in InfoBurst to query for any KPI’s that had not been entered by noon on the first day of the week.  This initiated an email Alert burst of a mini dashboard to the appropriate person responsible for updating the missing KPIs.  A second schedule ran 4 hrs later and emailed the Alert mini dashboard to BOTH the responsible party AND their manager!  This process ensured accountability for getting these Key Performance Indicators entered in a timely fashion for all the General Managers to be able to review in their meetings the following days.

I foresee exciting possibilities that this solution opens up for other customers as well.  Imagine being able to automatically email a dashboard to the VP of Finance when Sales Revenue  falls below a certain level, or ….  Think of the possibilities!

It’s not just that Xcelsius is going places globally but wherever it goes, it is creating new Business Intelligence applications that are slowly but surely changing the world.  

Of course, Rod Stewart is now in his late sixties and has been performing professionally for more than 40 years. He has always been a great entertainer and he knows his art to near perfection. He knows what to look and listen for in the people who support him and he knows what the audience likes and how to woo them.  His choice of songs from his incredible repertoire was spot on incorporating well known hits like “Maggie May”,  “Do you think I’m Sexy”,  and “Have I told you Lately” with some of his best songs like “Reason to Believe” ,  “The First Cut is the Deepest” and  “Hot Legs”.  His showmanship was in fine form too from kicking and heading soccer balls into the audience to getting down on all fours and worshipping the stunning stilettos of the very attractive female mandolin player!

The set lasted well over two hours but it was one of those concerts where you just did not want to end.  I left with a sense of euphoria and a desire to tell everyone about what I had just experienced (hence this blog).

I both attend and deliver a lot of Business Intelligence “shows”. To find that perfect combination of presentation, demonstration and connecting with your audience is not always easy. It takes practice, experience and, like Rod Stewart, a certain amount of showmanship.

Recently I had to deliver a presentation and demonstration of the complete SAP BusinessObjects product suite in one hour.  It is simply not possible to include everything in a meaningful way in that time so I focused  on the “well known hits” of Web Intelligence for reporting , ad-hoc query and analysis and Xcelsius for dashboards and “what if” scenarios. I had to explain the tools, demonstrate their usage as an end user and as a developer.

The customer was in the insurance business so I had one of my band members create fictitious data and scenarios for their specific industry. I also rehearsed the demo many times to make sure I could fit everything in the time allotted. I actually ran through it three times just before arriving at the customer site.

In the presentation room, everything was set up and just 5 minutes before the start, I plugged in a network cable and my Virtual Machine with the server software on it froze up!

Never happens – right?

Actually, it happens all too often. I quickly rebooted the Virtual Machine knowing it would take 10 to 15 minutes to fully reload and started the presentation on time showing my first PowerPoint slides while the server was rebooting in the background.  I managed to complete all the demonstrations within the hour and covered the Business Intelligence portal, Data Quality, Metadata Management and the semantic layer abstraction for good measure.

There were about 25 people present and I played to my audience making sure I had their attention and responding to their reactions as their eyes lit up with the data visualization in Xcelsius and the drag-and-drop magic in Web Intelligence.

They actually applauded at the end. I did not get the wild standing ovation that Rod Stewart got at the end of his set but then again, I did not have soccer balls to kick and stunning stilettos to worship!

Last week I received an email from the father of a girl who I used to coach on my soccer team. The email was entitled “Bragging Rights and Thank You” and he explained how his daughter had just been selected for the High School All Regions First Team. She is only a sophomore and helped her high school varsity team to win the State Championship in their division.

Coach Paul with the Desert Elite

I first coached her at our soccer club when she was 10 years old and she transitioned with me from recreational to competitive soccer. She was fast and fearless and it was not difficult to see that she had natural talent on the soccer field. During the 5 years I coached her, I encouraged her simply to do what she loved most which was to play well, be creative and have fun. I just watched from the sidelines and cheered as she improved and dazzled us all with new skills with each game.

Anyway, her father was justifiably proud and the “bragging rights” were in order.

So the day after, I receive an email from a customer who has been troubleshooting an Xcelsius dashboard issue with the help of one of our consultants, Roxanne.

Roxanne had suggested to the customer to create a copy of the dashboard and create new queries and XML cache’s with new names.  The customer was emailing to say that since doing this, he has been unable to recreate his original problem and is happy it is working.  He added a postscript that said,

“ I have to mention I used your name in naming the queries, xml caches etc while recreating. This must have had a positive effect”

 So it appears by naming the dashboard queries and XML caches “Roxanne” something, his problem is solved and his dashboard is running great! Well, I could not let this pass without both sharing it and bragging about it.

Roxanne is one of InfoSol’s top consultants and a true Xcelsius guru. Unlike my soccer player, she was a little more than 10 years-old when she learned BusinessObjects and Xcelsius but , just like my soccer player,  I simply encouraged her to do what she loved most which was to be creative and have fun. The result is a Business Intelligence Superstar whose mere name causes queries to behave and run better!!

If you want to determine if my bragging rights are justified, you can attend the Xcelsius  Master Hands-on Bootcamp at this year’s IBIS2011  which will be taught by Roxanne and witness first hand her amazing talents or you could just rename your dashboard queries after her!