Wednesday, December 19, 2007

Three Tips for Learning a Database

Rather than curse the darkness, Stella is going to light a little candle here. It seems that a lot of people never move beyond the basics of looking stuff up in a database: You need to find a person and you type his or her last name into the Last Name field and click find or submit or whatever. Congratulations! You are now working slightly more efficiently than if you had an index card file. The real power of any database is in finding groups of people who have things in common so that you can do things with them. For example, you could find everyone who had given to a particular area and then ask them to give more to a similar area. Or you could find everyone who lives in a certain area and then call them up to see if you can visit them. The possibilities are endless.

But if you're faced with an unfamiliar data structure, a completely horrible user interface, and no functional help files, how do you get from simple name look-ups to cooler stuff? Try these tips:

1. Don't be afraid. You're not going to break anything, really. In most databases you have to work pretty hard as a mere end user to mess things up, and so you'll be okay. If it asks you to save your changes and you either didn't intentionally change something or you're totally not supposed to change things, tell it to not save your changes. And if the database doesn't warn you before saving your changes, then it deserves what it gets and whoever manages the database should be taken out back, given twenty lashes with a slimy bit of pond scum, and then sent packing with his or her person still festooned with the aforementioned bit of pond scum.

2. Find a record that looks like what you are looking for. Let's say you know one of your constituents meets the profile of your target audience: he lives in Akron and has given more than $50 to the division of embarrassing conditions, and he's over 60 years old. Look him up by name and note where and how those other pieces of information (location, giving history, age) are stored in his record. Then head back to the advanced search interface and plug those criteria (location, giving history, age) into the right places, and suddenly you're finding groups of people who are like what you want.

3. Question your results. This may come as a shock to you, but databases are not magically created; rather, much like Soylent Green, it's peeeeeeeeeple. People make the data in your database. And people make mistakes. And store the wrong thing in the right field. Or store the right thing in the wrong field. And neglect to tell you that the advanced search features of their databases are terminally f*cked sub-optimal and cannot be relied upon. So anytime you're working with a new database or you're starting to push your knowledge of queries past the basics, it's important to go through your results with a fine-tooth comb to make sure that you have what you think you have. If it's a small, important group of results, I look at each one to verify that it does meet my criteria and that nothing unexpected has happened. If you have a really big result set, do some sampling to check it out. Look for problems -- someone who should not be on the list, at least on the face of it -- and then try to figure out how they got there. Unless your database is truly awful, there will be a reason that an out-lying record is there -- it's that people problem again.

With a little curiosity and some purposeful clicking, you'll be on your way to generating magical lists of people, places, and things to use in your work. And you can stop asking me to get you these lists. And that's good for everyone.

No comments: