[Various details redacted.]
I'm currently about to start a Geography degree at the University of [Redacted] at [Redacted] with a focus in GIS, and I've been finding that I have an interest in working with imagery. Obviously I should take Remote Sensing and other similar classes, but I'm the type of person who likes to self learn as well. So my question is this: What recommendations would you give to a student who is interested in working with imagery? Are there any self study paths that you could recommend?
I’m self-taught myself, and there are a lot of important topics in GIS that I don’t know anything about, so I can’t give comprehensive advice. I haven’t arrived anywhere; I’m just ten minutes ahead in the convoy we’re both in. Take these recommendations critically.
Find interesting people. You’ll learn a lot more from a great professor (or mentor, or friend, or tutorial) talking about something outside your specialty than you will from someone boring who’s working on exactly what you’re interested in. Don’t get insular! All the best artists I know have close scientist friends and vice versa.
That principle alone should expose you to enough interesting ideas that you will be able to see the most productive paths for yourself. I guess I could go on:
Look for real problems. “Let’s make a map of the furthest point from a McDonalds in each state” may be a useful exercise, but it’s not a real problem. Accurately measuring how earthquakes propagate is a real problem. Making sure that indigenous land rights are represented is a real problem. Finding early evidence of village destruction is a real problem. That doesn’t mean you have to spend all your time on scientific and humanitarian topics, especially as a student! But your work is valuable and should be spent on things you care about, even if they’re silly. If you learn to ask interesting questions that no one else is asking, you will get a good reputation among the people whom you would actually want to work for.
Learn as much statistics as you reasonably can. Trust me. Half the time I solve a tough technical problem it involves learning some stats, and then suddenly I see all these other places where that bit of knowledge applies. In fact, I’m making a note: I should learn more stats.
Read Edward Tufte’s books front to back several times, even the parts that don’t seem to have anything to do with maps.
For inspiration about image processing techniques in GIS, look at other fields: astronomical imaging, computational photography, knitting, archival photo restoration. I’ve been tinkering with some imagery taken by a satellite sensor that works such that the red, green, and blue channels are recorded at slightly different times, and having read about recovering early color images years ago turned out to be super valuable: I already know something about this problem that other people don’t! I mean, it’s obvious, but you find new and better ways of doing things because you know something that people in the field don’t already, so look at other fields.
Whenever you can, let the data lead. If you stumble on something interesting – say, some country you’ve never heard of releases a tranche of high-res imagery, or a major site opens up a set of anonymized geotagged data, or whatever – dive in, figure out how it works (technically and semantically), and see what you can discover.
If you need to do some work (a course, a job) that you don’t believe in, fine, and try to learn something from it. Pay the bills. But don’t bring it on yourself. Don’t say “Well, Yoyodyne makes kitten-seeking missile guidance computers and their contract doesn’t allow side projects, but I need something solid on my résumé, so I’ll just spend four years there while I get on my feet.” It might work by chance, but odds are it’ll lead to selling out, burning out, and just generally being no use to yourself or anyone else. The GIS industry is a moving target: don’t aim for a good job, aim to invent it.
One of my coworkers came to the company from a project to map the history of the ancient Mediterranean; another got on the radar because he was mapping the NYPD’s use of stop-and-frisk. Our lead developer was a philosophy/studio art major. I dropped out. Very few of us majored in GIS or CS. But we all ended up here because each in our own way we fundamentally care about geography – place, space, and helping people work with representations of their world more easily. Of course we’re a startup so things are a little unusual, but really, to any company worth working for, showing initiative, carefulness, curiosity, and delight in geography itself matters so much more than any one item on your CV.
When people say “do what you love” they don’t mean “goof off and trust the world to provide”; they mean “you’ll be working below your abilities whenever you don’t have intrinsic motivation, so find it”.
Stick with open-source tools as much as reasonably possible. There are various advantages, but one is that in principle you can always look inside them and figure out exactly what they do. In practice that’s rarely easy, but it’s still valuable. You also get to share your work with a much larger community – given that maybe 3% of the population can afford a closed-source GIS package, 97% of Earth’s latent GIS talent is in the open-source world. Help bring it to fruition:
Teach. Any time anyone is paying attention to you, you’re teaching anyway, so it’s good to be deliberate about it. This might take the form of a notebook blog, for example: “Today I tried to do X with method Y, but got result Z. Will try again next week”. Teaching forces you to think carefully in certain ways (as does programming!). Teaching also helps you keep ethics in mind. Mapping is a special kind of power that most people cannot tell is being abused even when it is; having to justify something to a student is one technique to keep in mind the consequences of things.
I think that’s enough general pontificating. Let me link to some stuff that might interest you:
– I use EarthExplorer pretty much every day including weekends – drop a pin and look at all the free data sources
– EOSDIS Worldview is a nice reminder that I live on a planet; see also MODIS Today
– Coverage of HOT OSM efforts in the Philippines after Haiyan/Yolanda
– King County (Seattle) is doing some quietly groundbreaking work on seeing geographical features like “treatments” in a standard epidemiology study, which gives them access to a set of statistical tools and approaches that are so far mostly used halfheartedly if at all in mainstream GIS
– James C. Scott works on how geography-as-we-know-it is used as a tool of power: I really recommend reading one of his books while in school, as an antidote to shallow triumphalism
– Tele-present water is some of the only geography art that seriously moves me
– Ingrid Burrington (@lifewinning on Twitter) makes a lot of the rest
– And James Bridle makes a lot of the rest after that, for example on submarine cables
– Peter Richardson’s “The Lay of the Land” is the only short prose I know that gets at the sense of place like it does
These are just things that are a little off the mainstream of GIS but have given me some kind of sustenance as a geo person.