"The Ten Commandments of Building Your Own Database"
1. Use the computer as a source. The most compelling stories can be put together with readily available information that merely needs to be organized and examined. Such records include standardized test scores available to every parent, traffic tickets, lawsuit records, public contracts, campaign finance reports, city/county school budgets, personnel files, public official voting records, billing records, check files.
2. No amount of data is too small an amount. How big does a story have to be before you plug data into a spreadsheet or database? The answer: "smaller than you'd think." Reporters love to say, "We analyzed a gazillion records and found" That is well and good, but you can add meat to almost any story - and improve chances of landing on Page One - by plugging a dozen records into a spreadsheet.
3. Don't be discouraged if you have to do it the hard way. Sure, it's great when there is a ready-made database waiting to be analyzed for investigative stories and a cooperative public official willing to share it. That's not always the case. Besides, if the agency keeps records on a database, the target is already ahead of you. Chances are, if you're the first person who ever took a look at the numbers, there's a good story to tell. If the agency can't analyze its own data, it has no idea what's really going on.
4. Think about the best way to get data before you go and get it. A low profile works best when you're taking on a new target. You don't want to give them time to circle the wagons and take action that would kill your story. So, find out if they release the data to the general public in some fashion and then get those reports acting as Joe Citizen. It's more work, but no one will know what you are up to.
5. Organize your data before you analyze it. Think about the questions you want your data to answer before you start working on your spreadsheet, and then customize the columns to fit your needs. Think in terms of yes/no questions. Find ways to create numbers that will allow you to rank your data.
6. The "sort function" is your friend. The sort function may sound boring, but it isn't. It is the most powerful key on the computer when it comes to CAR. Do you want to find out who made the biggest campaign contributions? Or which school had the largest test-score swings? By sorting the data, you'll often find your most powerful anecdotes.
7. Keep your data consistent. If you're doing a column of people's names, always put the last name first. Don't use 9 and nine. Don't do anything that will make it harder to rank or sort data or to find data that matches.
8. Double-check the data. You need to be bullet proof. Mistakes are deadly. They give your target a chance to discredit your research.
9. Do it yourself whenever possible. Sometimes you have no choice about who inputs the data: there is simply too much to handle by yourself or with another reporter. But whenever possible, type in the data yourself. Two advantages: you see stuff you wouldn't have found otherwise, and you eliminate the errors that other people with much less invested in the story are going to make.
10. Buy yourself enough time. Editors think CAR is easy. It's your job to make sure they know you're going to spend two weeks doing nothing but punching numbers into a spreadsheet. And then maybe another week checking your work and playing with the data before you're ready to make a single phone call. Tell them up front. You need to sell them on WHY there's no other way to do it.