8,200 Radiocarbon Dates and Counting . . .

4/18/2016

At the risk of alienating the few friends I have, I'm going to admit that I find working with large datasets to be a rewarding pastime. I'm not sure why, but I feel like that there is something really satisfying about compiling, cleaning, standardizing, and finding and filling gaps in data. So I've gone from the "why doesn't this exist" stage about a month ago to now having a dataset of over 8000 radiocarbon dates pegged to county-level provenience. Here's the current map:

You'll notice that several areas have been filled in since last week's map. Matt Boulanger sent me a large spreadsheet of dates from New England and the Northeast, most of which I've now incorporated (I still need to track down counties for some of the dates). Victor Thompson and Matthew Colvin sent me a spreadsheet of the Florida radiocarbon database assembled by Steve Dasovich and Glen Doran, so those dates are in the database now also. I've begun going through the chapters of the 2009 Archaic Societies volume (edited by Thomas Emerson, Dale McElrath, and Anderw Fortier), which contain significant listings of Archaic Period dates from the Midcontinent. I know of compilations from North Carolina and Mississippi - those will be coming up.

I've got a few dates from Ontario and Quebec, but I confess to being a bit mystified as to how the various administrative regions, counties, and census divisions of those provinces correspond to the provenience information I have for the dates and the GIS data I'm seeing in front of me. Nova Scotia and New Brunswick were easy, but I will need to figure out what's going on in the other eastern provinces of Canada before I can display the data.

Assembling and formatting a bibliography to go along with this dataset will be an unholy nightmare. I'm going to need to get that done before I release the data in a usable form. Working on pulling all the citations together in one place and correctly referencing them in the database may be a good summer job for my daughter. As far as I know, she doesn't read my blog. Maybe that's for the best.

Update (5/4/2016): 9,129 dates and counting . . .

Cameron Howell

4/18/2016 06:03:52 pm

Hey very cool. I was working on a similar project a few years ago also set up at the county level. My goal was to make an online accessible database similar to the CARD database for Canadian dates. I have about 5.5k dates, and I would love collaborate to help put these kinds of big data sets out there.

Andy White

4/19/2016 06:59:28 am

Sounds good - send me an email? [email protected]

Jennifer Birch

4/19/2016 08:27:55 am

Yes, I was also going to make you aware of CARD: http://www.canadianarchaeology.ca/

Andy White

4/19/2016 08:30:05 am

I'm aware of it, thanks. I've been using it to try to track down missing bits of information and resolve inconsistencies.

Bob Jase

4/19/2016 08:30:32 am

I was a bit flummoxed when the previous map showed zero up my way, now I can relax.

Andy White

4/19/2016 08:31:50 am

It's a work in progress, Bob. At this point absence of evidence is not evidence of absence.

pferk

4/21/2016 03:58:15 pm

Me too ... am making a database of finds at over 1400 sites... get to listen to much good music whilst assorting

Andy White

4/26/2016 06:46:08 am

I'm on a Suzi Quatro kick right now.

Ron Halliday

4/25/2016 08:59:10 pm

How detailed is your original data? A few years ago I developed a map that clearly and consicely shows seismic event point data by size, depth, and precision. It might be an interesting exercise to apply some of the same techniques to your data (applying how I visualized ranges of epicenter accuracy to ranges of carbon dating accuracy, for example).

Ron Halliday

4/25/2016 09:05:15 pm

As you probably know, typical earthquake maps only show magnitude, epicentrer and depth; my map's points also included epicenter precision and magnitude calculation details (measured vs estimated, for older events).

Andy White

4/26/2016 06:51:41 am

Fundamentally, each piece of data will have a spatial reference (county level), the measured number of radiocarbon years before present (RCYBP), and the 1 sigma error term. For my own purposes, my goal in assembling all this is to be able to use radiocarbon data as one line of evidence for evaluating ideas about population dynamics during the Late Pleistocene and Early Holocene. The first thing I'll probably do is "clean" the dataset I use for analysis by ignoring dates with large errors (say, over 150 radiocarbon years). But it will be fun to play around with the whole dataset in GIS once I'm too that point. And I'll make it available so everyone else can use it, as well.

MKPolutta

5/2/2016 08:50:09 am

As a librarian and techy type, I thought I'd make sure you were aware of Open Refine, which is software that is great for data clean up. See http://openrefine.org/ It's FOSS, and is definitely being actively worked on.

Andy White

5/7/2016 04:27:32 am

I'm aware of it, but haven't used it extensively. We're cooking up a scheme to publish a database, so we'll need to be working on it together and collaborating remotely on sorting out citations.

Bill Whittaker

5/20/2016 07:51:08 pm

Let me know when you get to Iowa, I can help.

Andy White

5/26/2016 03:23:58 pm

Thanks - will do!

Comments are closed.

8,200 Radiocarbon Dates and Counting . . .

Archives

Categories