Calling all GIS-experts in da house

September 3, 2007 1:23 AM

Brain storm-fodder: What's the most innovative idea you've got for an thesis on GIS (Geological Information Systems) and archeology?

An old friend is having trouble to find a suitable idea for an undergrad-thesis on GIS (geological information system) and archeology. We're holding an brain storm-session in a while and I'd like to be able to throw him some meat-bones to chew on (mentally speaking).

So, calling all GIS-experts, archeologist and who ever would like to chip in: What's the most innovative idea you have for using GIS and archeology that's suitable to explore in an under-grad thesis?

(The googling reveals several master thesis about archeo-GIS, but they're on a too high level - academical and resource-speaking). I'm looking for something somewhat "lighter". My apologies if the question is to broad.)

posted by Rabarberofficer to education (9 comments total)

Boy, this is an wide-open question, especially since we're talking about an undergrad level thesis. In my opinion, your friend needs to really take a narrow approach. Figure out a geographical area, a technology, a single site, etc. In particular, for undergrad level theses it seems like you can almost never go wrong with taking an approach that has been done in a different geographical region and applying it to a new one. For example, if GIS has been used to track early village sites in the Missouri basin, see if there is suitable data to do the same thing around Phoenix. I may be overly cynical, but the stories I've heard about undergrads flaming out at the thesis level are almost always people who have tried to have projects that are new and innovative and who forgot about having projects that are doable.

I'm assuming that since this is undergrad level, your friend won't be doing his own data collection, so which datasets does he have access to? In terms of possible topic brainstorming, what is his geographical or technological area of focus? Heck, where is he located?
posted by barnacles at 1:36 AM on September 3, 2007

I'll second that your friend needs to find a specific focus and not try and do something too big. And as barnacles asked above, what are your available datasets? I dont know where you and your friend are located, but if you're in the US, does the university have CRM-type reports for projects/excavations/field schools they may have done in the past? More specifically, does he have access to that data? That would be a ready made dataset to start playing with. For undergrad work, taking some field data and doing a simple distributional model is probably sufficient, assuming that all one is looking for is an example of GIS application to archaeology rather than actually trying address a specific research question.

I'd suggest staying away from "innovative" and keep it simple. Leave the heavy stuff for a master's thesis.
posted by elendil71 at 3:55 AM on September 3, 2007

Nthing the keep it simple theme. A lot of people rush into theses with the idea that it will change their field or turn part of their specialty upside down without realizing that undergraduate, and to a good extent, M.S., theses are to do the scientific "gruntwork" that needs to be done in every field.

That isn't answering your question, though. With that in mind, the best and easiest thing your friend could do is read papers about GIS-archeaology in the area/which he has access to data-sets and look for "this was not done at this time" sort of sentences, because most papers mention one or two things that still need work. Also, your friend should be working with a professor who should be helping him and who probably has a pesky little data-set he/she would just LOVE to have someone do some mind-boring tasks on it.
posted by barchan at 5:42 AM on September 3, 2007

Coming from a background in archaeology and geography, and having completed one of these GIS-in-archaeology theses, I have to cynically say that the bar is currently very low. Very very low. Though GIS software has become commonplace in archaeological projects, it is primarily used as a basic record of spatial information, barely different than a paper map or technical drawing. Most archaeological projects that actually use GIS analysis techniques are basic point-plotting with statistical characterization of a few basic attributes... for example, finding the most typical ranges of a few variables that characterize a majority of Native American settlements (e.g. distance to water, slope, aspect, elevation). There are notable exceptions (which I'd gladly discuss further), but most archaeologists using GIS are not being innovative--they are making site-location maps.

HOWEVER, it does not need to be this way, and the best thing one can do to expand the range of possibilities is to look for examples in other disciplines. Species distribution modeling, of both flora and fauna, has been rapidly developing in the last few years. Doing the same rote analysis I ridiculed above, but with a comparison of the best algorithms/methods would already be paving new ground.

One aspect to consider is whether the goal of this thesis is to write about GIS within an archaeological context or to use GIS to write about archaeology... If it is the latter, there is a time and place to contextualize the research. If it is the former, you can choose the time/place to best fit the method. And there is so so so much previously-collected data which could be re-analyzed with "modern methods."

I have several colleagues who are very active at this GIS/archaeology interface and would be glad to pass along information via email.
posted by zachxman at 6:41 AM on September 3, 2007 [1 favorite]

I would ask your friend what his future goals are: working with GIS or archaeology? If the answer is working with GIS he needs a thesis that demonstrates his capability with GIS work. If he wants to work in archaeology he needs to find a thesis that answers a legitimate archaeological question that researchers want to know. In archaeology he would probably want to study in his area of specificity so he can apply to graduate school. I would also echo Zachxman that he needs to use the GIS to full capability. You don't want your prof to look at the assignment and say, well, I could have done that on paper just as easily.
posted by aetg at 7:10 AM on September 3, 2007

Thanks for all the great input!

I'm sorry that I've failed to give adequat details. I'll certainly put FOCUS in capital letters on top of the list for my friend.

Zachxman and Aetg have given the most useful tips for the brain storm with my friend.

Brief details about my friend: He is a trained archeologist and have been doing GIS-work in that field. BUT he has not yet made any academic viable work. It's all hands-on, and no paperwork, as far as I can see.

So, he's quite knowledgable and needs to find a topic that will a) make fair use of his knowledge and b) be useful for work-applications in the archeo-GIS-field.

For what it's worth, he's stationed in Sweden (vikings, yes I know..)

Again - thank's for your input.
posted by Rabarberofficer at 8:40 AM on September 3, 2007

GIS can work at any scale of course, and archaeological studies exist at many scales as well. I've seen GIS technology used to create Digital Elevation Models of single stone tools to look at such attributes as use-wear patterns, and also at the scale of entire continents. I'd nth the advice to choose a small topic for an undergrad. The learning curve is steep on GIS but if he has some knowledge of that already then, as barnacles says, the bar is set very low as most GIS applications in archaeology have been quite trivial.

So, anyway, here is my idea: archaeologists very commonly look for clusters in their data, and spatial clustering is usually treated as a simple derivative of distance - that is, the clusters sought tend to be "globular" in some sense, even those that use a friction surface or other derivative of topography. In Sweden there are a lot of culturally modified trees (see the witness tree thread, up-page). In BC, I have seen one study which I thought was really innovative (but not GIS) which interpreted a distribution of stone tools at a single site as reflecting discard of stone tools alongside logs that were being worked on. This kind of "linear cluster" was not at all apparent in any kind of spatial analysis and many archaeologists were not thinking of trees as a structuring agent in their sites. It turns out that GIS has some powerful capabilities to recognize linearities in the data, and these might be useful in identifying patterns in archaeological sites or even regionally - maybe to identify logs as structuring features, or any other number of things.

OK that was a bit long.

There are also a lot of potentially cool projects that could be done with google earth, a GIS, and some archaeological data. Some archaeologists are using google-earth to integrate their data into regional landscape visualizations - e.g., Michael Blake at UBC. This would potentially give your friend good skills in GIS, web-interface, and public archaeology.
posted by Rumple at 10:13 AM on September 3, 2007

GIS only really comes into its own with really large datasets. 90% of Arc* is about spatial processing and display, not just a few points. If you have a large dataset, I'd avoid just mining it for any random correlation that happens to be there - even significant correlations can (and do) appear randomly, and you're tempting fate by going on a goose chase without a single one picked out.

I'd be inclined to try to spatially link it to some available modern-day dataset. The only hard part with that is to avoid bias because of dig sites, so you'd need to be careful about making sure you'd be testing what you think you are. Take a browse through ArcCatalog one of these days though - there's a surprising amount of random stuff in there. Climate datasets (especially very specific ones) would be one place to start. Maybe pop density, some habitability score, soil quality, LULC classes, geological faults. You'll need big datasets & a good thesis, but GIS could really support something like that.
posted by devilsbrigade at 12:16 PM on September 3, 2007

To follow up with Rumple, take a look at the Minnesota MapServer if you're doing web GIS stuff. There are prettier ones around, but MMS is standard and straight forward, and makes a good free alternative to ArcWeb.
posted by devilsbrigade at 12:19 PM on September 3, 2007