Saturday, November 1, 2014

GIS for Onomasts

http://onomastics.co.uk/gis-for-onomasts/

GIS for Onomasts: Report from the Advancing Onomastics workshop 28—30 October 2014
Rebecca Gregory, PhD student, School of English, University of Nottingham
PhD students in onomastics were lucky enough this month to have the opportunity to attend a training course, organised by the Universities of Glasgow and Copenhagen and run by Peder Gammeltoft, on QGIS software and its use in an onomastic research setting. For an explanation of what GIS is, and its possible uses for name-scholars, take a look at the March blog by Peder Gammeltoft.
At the last Advancing Onomastics workshop in Copenhagen, participants expressed a wish to learn more about GIS software. Many training courses in the use of different GIS software packages are tailored to researchers and professionals whose data is very different from that used by name-researchers. It was suggested that far more use could be made of GIS by PhD students if they were taught how to employ it in ways more relevant to their research. There are two main uses for researchers: analysis of data within their work, and presentation of their findings to others. There are some functions which it would be either very difficult, time-consuming, or sometimes impossible to achieve without the use of GIS.
The desire to learn more about GIS is apparent in both the number of students who attended the workshop and the distance they travelled to be there! Participants included many students from the UK, but also those working in Italy, Estonia, Denmark and Norway. Areas of interest included (but were by no means limited to) personal, settlement, minor and botanical names, and researchers had links with disciplines as wide-ranging as archaeology, literature, genetics and sociolinguistics.
The workshop itself took place over three days at the University of Glasgow. The course was divided into sessions, each of which had a clear aim and structure, with materials designed to give both theoretical and practical instruction, as well as opportunities to incorporate participants’ own interests. The course took place as follows:
Session 1.1: General introduction to GIS
In this opening session, we learnt about the uses of GIS and GIS software: in the most simple terms, the viewing and editing of digital maps, the creation of spatial data, and the production of custom printed (or otherwise presented) maps. We discussed the types of data used by GIS software, and where this data, both official and unofficial, is available from.
Session 1.2: Introducing QGIS, learning the basics
QGIS is an open-source application, meaning that it is free to download and is frequently modified and updated by multiple developers. This has its advantages (e.g. the ability to use file types associated with other software; online manuals and advice available), but also its disadvantages (it can sometimes be less polished and more prone to bugs than some other software). There are, of course, many other software options available for GIS.
We familiarised ourselves with the QGIS interface and made sure we had all the features enabled that we would need. We then learnt about GIS data and how to open files from different sources, and to view the “associated information” which relates to the geographical points or areas (things like place-names, parish names, feature types) within the program. We then experimented with ways to display information, changing colours and symbols, adding labels, and in some cases creating some rather psychedelic maps!
blog4
The result of one student’s playing with the display options.
Session 2.1: Creating GIS layers in QGIS
Now we had learnt how to load and style maps from outside sources, we were taught how to create our own maps. As well as modifying and merging existing data, it is possible to create your own maps from scratch. Some ways that we use GIS are for displaying onomastic information (e.g. distribution maps), and spatial analysis (altitude, slope, distance etc.): this is the kind of thing which is incredibly difficult and time-consuming if done without the assistance of computer software.
We learnt how to mark, label and categorise points on a pre-existing map, to calculate their coordinates, and to save this information for later use. As well as points, we created layers for lines (such as streams or roads) and polygons (lakes, parishes etc.).
Session 2.2: Georeferencing digitised paper maps
Much onomastic research uses documents or attestations from before the time of digital mapping. While some older maps have been made available for use in GIS software by organisations such as the Ordnance Survey, many more have not. It is possible, however, to scan paper maps and to extract data from them. We discussed the kinds of maps which can and cannot be georeferenced: a map which is not reasonably accurate cannot be used in this way (see below for a perfect example of a map which should not be georeferenced).
blog10
If accuracy is the goal, this is not the best map to choose for georeferencing!
Our task was to georeference a map sheet of southern Shetland, using distinctive points on the map to anchor it to the same points on a map which was already georeferenced. Once these “control points” were marked, the program did the hard work for us, resulting in some very accurately georeferenced maps.
blog2
The old map sheet georeferenced to fit with our Shetland map.
blog1
A close-up of our Shetland map (transparent green) with the georeferenced map sheet underneath it.
Session 2.3: Creating new datasets from existing and new sources
In this session, we adapted a map of parish boundaries in southern Shetland in 1930 so that it showed the 1878 parish boundaries. To do this we had to divide some of the parishes along additional lines as marked on the map we had georeferenced in the previous session. In other cases we had to combine two or more parishes where there should be only one. This task began to show us how some relatively simple steps with GIS software can be used to make huge and useful changes to make data more relevant to different time periods and foci.
Screen Shot 2014-10-31 at 21.48.35
Using the georeferenced map sheet and the QGIS polygon splitting tool to draw the older parish boundaries.
parishes1930
Pre-splitting: the Shetland parishes as they were in 1930.
Shetland Parishes 1878_1
After splitting: the Shetland parishes as they were in 1878.
Session 3.1: Analysing onomastic data in QGIS
Naturally the title of this session could have included enough information to keep us busy for weeks; however, Peder Gammeltoft chose some simple methods of analysing and displaying data which were most likely to be useful to all the participants at the workshop. We learnt how to display areas (in this case parishes) by colour; to calculate and display (by colour) numbers of a particular name or element occurring in a parish or other geographical division; and how to create a heatmap. We also learnt to create different backgrounds on which to display our points, lines and labels (see below for an example background).
Screen Shot 2014-10-31 at 21.50.55
A heatmap showing place-name distribution in Shetland
Screen Shot 2014-10-31 at 21.52.28
Part of Great Britain and Ireland with the Hillshade effect applied
Session 3.2: Presenting onomastic GIS data
This final session concentrated on how to present the data we work with in our research. This could include, of course, use in slides for conference papers, or as illustrations for a book or article. QGIS includes a “Print Composer” which allows very precise arrangement of map layers, scales, legends, labels and other features to create the best output for your intended purpose. In this way, illustrations can be tailored without the data layers themselves having to be re-created each time.
On this course, all the participants received a level of training in the use of QGIS software which was specifically designed to facilitate its use in onomastic research. I feel confident in speaking for all the participants when I say that I am more than a little excited about the possibilities for my future research using GIS. Our thanks go to Peder Gammeltoft for running the course, and to the other “behind the scenes” support at the Universities of Copenhagen and Glasgow. Many of the images in this blog are from the course material, and credit goes to Peder for them.