For 2007, The Password Is...

Happy New Year to all...

I'm probably dating myself here - well, OK, I'm definitely dating myself - but I can hear the hushed tones of the announcer on the old Password show right now...

"The Password is...'Context'"

As I get farther and farther down the rabbit hole of Master Data Management, Metadata Management, Data Quality, et al, everything seems to come back to that word - "context". Dictionary.com defines "context" this way:

1. The parts of a written or spoken statement that precede or follow a specific word or passage, usually influencing its meaning or effect: You have misinterpreted my remark because you took it out of context.

2. The set of circumstances or facts that surround a particular event, situation, etc.

(I left out the other meaning dealing with the structure of a mushroom. You're welcome.)

In my opinion, the thing that separates information from data is context. A manager really can't analyze the effect of a particular event or group of events if she does not understand how these events relate to other events, either preceeding or succeeding. When the definition of "context" is applied to data, you can see that each piece of data is related to every other piece of data either directly or indirectly, and the nature of these relationships give the data character and relevance.

One way of providing context in the business intelligence world is by providing dimensions to classify events. Dimensions are the attributes that describe events, or facts. Generally, these are presented as "sorts" in a reporting tool on such topics as Department, Business Unit, Time, and Account, to name a few. They are used to perform the infamous "slice and dice" operation to information to determine trends and seek outlier data, and they do a good job of it for the most part.

However, I feel that the way dimensions are usually used is very limiting. The definitions of these relationships are static and linear, which doesn't reflect reality very well. It reminds me of why I didn't become an Economics professor - I didn't think that the theory of ceteris paribus* held water and I couldn't stand the thought of teaching that to others while a better model was (hopefully) being discovered.

Over the next few months, we'll be exploring the concept of context in data warehousing and how to better represent it with the tools we have (and maybe stumble across a new technique or two).

...and yes, I plan to finish the "What's Wrong with Process" eventually as well...I'm still working on that writer's discipline...

* "Cetris paribus" is defined as "With all other factors or things remaining the same." I remember all of this from Econ class, but I had to Google it to get the right term 20-plus years later. Thanks again to the power of the Internet...;-)