“A measurable value, often numeric and typically aggregatable, stored in a data warehouse.
A schema object representing a column in a data warehouse table and containing basic or aggregated numbers-usually prices, or sales in dollars, or inventory quantities in counts.”
How…technical.
Let’s explain it in terms the terms of an “every day life” transaction. If you bought shoes at the shoe store, the transaction involved a certain amount of “facts” and “attributes”. The facts in this case are “dollars, and quantity”. You gave something measurable, to get something measurable. Your fact is the base level of the values in the transaction that we later turn into metrics. The attributes in the transaction are almost limitless: the Converse All Stars, the shoe size, the store name, the person who you bought the shoes from, the time, day, and month that you bought them in. Metric are then built around facts. The metrics would be total dollars spent, average sales price, or the number of sales.
Okay, so this is an old one, but I still find it useful.
If you can’t run a report because of a spool space error, and you can’t convince the DBA to give you a couple of gigabytes more, then you might want to go to the Report Editor, Data ->VLDB Properties -> Tables -> Intermediate Table Type and select True temporary table.
What this does is modifying the SQL so that instead of having a huge chunk of code (that would use up a lot of memory) it breaks the code into CREATE – INSERT bits that are far easier to handle. It also makes the SQL more readable by humans (this includes programmers).
By the way, this setting also comes to rescue when you get the “Database is terrified by the script and it can’t even begin to analyze it” error. I don’t have the specific error message at hand, but I know for sure that Teradata does this when the SQL code is all in one pass and is so large that it can be seen from the Moon.
Here’s the conclusion, click through to see all the detailed points.
In essence, a reasonable business case will demonstrate that the potential financial improvement is offset against the increased level of effort and investment in putting the MDM (Master Data Management) program in place. This business case shows that the benefits outweigh the costs and are achievable (and measurable) within a reasonable time frame. Those benefits may be financial, or may reflect other key performance indicators. But the absence of a business-oriented value proposition and the inability to socialize that business case will only lead to continual questions about the need for MDM.
When I’m creating a report, and adding elements to the data warehouse, I need to make sure that I understand why this data is important, and what it’s going to be used for. I’ve had many back and forth conversations between people who believe in the “just give them what they’re asking for” approach, and the “understand what they’re asking for” approach. I’m obviously in the second camp. With Business Intelligence, I believe that it’s important to understand what the end game is with regards to the data that you’re capturing. It’s one thing to just spit out reports, but many end users aren’t aware of what solutions products like MicroStrategy can offer. It’s our job to offer alternatives, to improve the workflow process, and to understand what the data is going to be used for…otherwise we’re not offering strategies and solutions, as much as we are migrating current processes to a new platform. That just seems like wasted dollars to me.
Business Intelligence is just that, Intelligence. If we’re in positions where we’re simply recreating spreadsheets, we are ignoring the opportunities that true dynamic data has to offer, and failing to give our feed back as people who understand what software solutions can offer…then we’re not living up to our potential. For many of us, it’s our job to push back, ask questions, understand why people want what they want, and offer them the myriad of solutions that true BI has to offer. This doesn’t always make friends…people get married to processes, and get stuck in the “now now now”; many don’t take the time to step back, and examine how the data processes can be altered and improved.
Here’s some weekend reading suggestions for any of you data heads out there. Everything from the Data Visualization essentials from Tufte and Few, to a few books about Data Warehousing. Great Stuff:
There’s really is a lot that can be learned about Dashboarding through the reading of these books, and check the sidebar for the link to Perceptual Edge. I have to give credit to one of the better presentations I attended that at MicroStrategy World 2009 by Michael Sandberg, who at the time was with Disney, and turned me on to some of these publications.
Recent Comments