My career has covered a lot of ground. I'm a decent programmer, a solid sysadmin, I've worked with database and network administration and so forth, and I'm pretty good at all of it. But there's one thing that I feel I have a genuine talent for, and I've only rarely had the chance to do it: data analysis.

For example, at one job, we were asked to make a documentation archive available online for the first time--30+ years of scanned PDFs covering thousands of products in 30-odd different languages; they all had hand-edited XML metadata, but it was often incomplete, erroneous, or used the wrong encoding.

My team lead--a better programmer than I was--examined it and decided we'd have to examine every document manually, at a cost of some 200 person-hours. I asked for a few hours to try to pick out just the most complete entries. In the process of writing that script, I noticed that many of the errors repeated consistently, and I could deduce and correct them from a few data points. I put in three 12-hour days, iterating and examining and iterating again, finding and characterizing patterns, and by the end of it I had a script that correctly classified 98% of our documents.

Taking a big, sloppy dataset and teasing out the clean patterns is enthralling to me--I literally do this for fun, with videogame data--and I seem to be good at it despite zero formal training. But just saying that doesn't impress interviewers, and I really don't even know what kind of jobs I should be looking for. I've been calling it "data analysis", but Microsoft apparently calls it "data science" or "data engineering," depending on the details.

So I guess my questions are: How do I find work like this? What should I do to market myself? Are there courses I should take to teach me standards, methods, and tools that I'm not aware of, and/or certifications that would appeal to employers?

Incidentally, I'm going to be looking for work in the Seattle/Redmond area soon.