I’ve been thinking a lot about archaeological data and its transformation, whether in the field or during ‘post-field’ analysis. More often than not, we are transforming spatial, temporal and thematic information.

More recently, as I’ve been working with records that someone else has made, I keep coming back to “how much do I trust this source”? Or more precisely, how much do I trust the spatial, temporal and thematic values that I have added to my spreadsheet? And how do I tell someone who isn’t familiar with the source, with archaeology, and with academic writing what is trustworthy and what isn’t? Can I visually present this information?

Let me explain by example. Since my last post, I’ve geocoded the data in the csv. I went with Chieko Maene’s excellent geocoding python script, available here. It calls arcgis and google APIs (with the option to add API keys from others), grabs the coordinates, and creates a new csv with the geocoder name. I chose this script because I wanted to know what exactly the geographic coordinates refer to – the place name, or the centre of the district, or something else.

Maene’s script returns coordinates for each place name, and also what the location refers to, such as the ‘building name’, ‘street address’, point of interest, administrative unit, etc. Mind you this doesn’t mean that building name is the most accurate and administrative unit, the least accurate. It just means that the geographic coordinate I now have refers to that ‘thing’. The ‘thing’ is not always the place name I’m interested in. In some cases, I have coordinates for buildings and roads that share a name with the place name. Not awesome, but at least I know this ie. I have a way of assessing its accuracy.

But what about the temporal information in the csv? How much do I trust periodization and the culture-historical period? I’m not so much concerned here with assigning each place name a date or date range, a task that has its own challenges. Rather, I am referring to the material culture that was analysed and upon which the archaeologist assigned the date/date range. Was it based on the pottery, stone/metal/bone tools, other artifacts? Was it on radiocarbon dating, and what was the context of the analysed material? This sort of information (in the source I am working with) isn’t always available or explicit enough to assess reliability. It can be straightforward, so “Harappan” might refer to pottery (although the term can be used as a ‘trait’), and not so easily interpreted, “Black-Red-Ware with lithics”. It is one thing to simply take what’s written and faithfully put it into a spreadsheet. It is another to encode information meaningfully.

Then of course, there is thematic information and the many different ways of describing a thing. A cromlech is different from a cairn, but the criteria isn’t clear (to me). So there’s quite a bit of interpretation going on in just a simple csv. Tags might be a way around this. But is it possible to present these metadata (imperfections and well, especially the imperfections!), visually? It sure would be nice.

In the meantime, I’ve started work in Mapbox. And as of last week, I have Web hosting, courtesy of the Matrix lab. My next post will focus on the Webmap, continuing work on map interactivity and developing the website to house it.