What is Project Crescent? The Microsoft SQL Server team blog describes Project Crescent as a ‘stunning new data visualisation experience’ aimed at business users, by leveraging the ‘self-service business intelligence’ features available in PowerPivot. The idea is to allow business users to serve themselves to the data by interacting, exploring and having fun with it. The concept at the heart of Project Crescent is that “Data is where the business lives!” (Ted Kummert), so business users have access to the data directly.
For many users, this new methodology of data visualisation could be a real fundamental change in their way of looking at data, a real Kuhnian paradigm shift; instead of using basic reports, instead accessing a self-service way of understanding their data without a reliance on IT, and without invoking a waterfall methodology to get the data that they require in order to make strategic decisions.
What does this data visualisation actually mean for business users, however? Haven’t business users already got their data, in the form of Excel, Reporting Services, and other front-end reporting tools? The answer to this question really depends on the particular reporting and data analysis process deployed by the business users. Let’s use an analogy to explain this. In the ‘Discworld’ series of books’ by Terry Pratchett, one book called ‘Mort’ contains a joke about the creation of the Discworld being similar to creating a pizza. In other words, the Creator only intended to create night and day, but got carried away by adding in the sea, birds, animals and so on; thus, the final outcome was far beyond the initial plan. The book continues that the process was similar to making a pizza, whereby the initial outcome was only intended to be ‘cheese and tomato’ but the creator ends up impulsively adding in all sorts of additional toppings. Thus, the final result is something that was over and above the original intention. Similarly, reporting and data analysis can be analogous to this process, whereby the original planned outcome is surpassed by the addition of new findings and extrapolations that were not originally anticipated.
Put another way, there are two main ways of interacting with data via reporting; one is structured reporting, and the other is unstructured data analysis. In the first ‘structured’ route, the report is used to answer business questions such as ‘what were my sales last quarter?’ or ‘how many complaint calls did I receive?’ Here, the report takes the business user down a particular route in order to answer a specific question. This process is the most commonly used in reporting, and forms the basis of many strategic decisions. If this was Discworld, this is your base ‘cheese and tomato’ pizza!
On the other hand, unstructured data analysis allows the business user to take a look and explore the data without a preconceived business question in their heads. This allows the data to tell its own story, using empirical evidence based on the data, rather than using pre-conceived ideas to generate the data. In our ‘Discworld’ analogy, this would be the final ‘toppings and all’ pizza, that contained so much more than the original intention.
So, Project Crescent is great news for business users for a number of reasons:
– users will be able to use ‘self-service’ to create their own reports, with no reliance on IT staff
– users will be able do ad-hoc analysis on their data without being taken down a particular road by a structured report
– the traditional ‘waterfall’ methodology of producing reports can be more easily replaced with an agile business intelligence methodology, since prototypes can be built quickly and then revised if they do not answer the business question.
At the time of writing, it is understood that the data visualisation aspect of Project Crescent will involve advanced charting, grids and tables. The users will be able to ‘mash up’ their data in order to visualise the patterns and outliers that are hidden in the data. Although it is difficult to quantify for a business case, it is interesting to note the contributions that visualisation has made to understanding data – or even occluding it, in some cases. One example is in the discovery of DNA: Rosalind Franklin’s photographs of the DNA structure revealed the double helix, which was examined and found by Crick and Watson. This has had enormous contributions of this finding to our understanding of science. On the other hand, incorrect data visualisation has been proposed as a contributor to the decision making processes potentially leading to the Challenger disaster by Edward Tufte.
So far, it sounds like a complete ‘win’ for the business users. However, it may be a Kuhnian ‘paradigm shift’ in a negative way for some users, in particular for those people who rely on intuition rather than empirical, data-aware attitudes to make strategy decisions. In other words, now that the ‘self-service’ Business Intelligence facilities of PowerPivot and Project Crescent are available, business users may find that they need to become more data-oriented when making assertions about the business. This ‘data-focused’ attitude will be more difficult for business users who use a ‘gut feel’, or intuition, to make their assertions about the business. This is particularly the case where business users have been with a company for a long time, and have a great deal of domain knowledge.
It is also important to understand that the ‘base’ reporting function is still crucial, and no business can function without the basic reporting functionality. Thus, Reporting Services, whether facilitated through Sharepoint or ‘native’, along other reporting tools, will still be an essential part of the business intelligence owned by enterprises. If this is Discworld, this would be our ‘cheese and tomato’ pizza. Put another way, there would be no pizza if there wasn’t for the base!
Terry Pratchett commented a while ago that ‘he would be more enthusiastic about encouraging thinking outside the box when there’s evidence of any thinking going on inside it’. This is very true of business intelligence systems, as well as the processes of creative writing. The underlying data needs to be robust, integrated and correct. If not, then it will be more difficult for business users to make use of their data, regardless of the reporting tool that is being used. In other words, the thinking ‘inside’ the box needs to be put in place before the ‘out of the box’ data analysis and visualisation can take place.
Project Crescent will be released as a part of SQL Server Denali, and will be mobilised by Sharepoint and the Business Intelligence Semantic Model (BISM). A final release date for SQL Server Denali is still being negotiated, but I hope to see it in 2011. To summarise, Project Crescent offers a new way of visualising and interacting with data, with an emphasis on self-service – as long as there is thinking inside the box, of course, regardless of whether you live in Discworld or not!