metal detecting debates

Tinkering with the machine and linking data

This post is rather belated, I’ve had a lot of things on over the last week. Family, server hardware problems, filming a short make believe piece for a children’s video conferencing workshop, editing and publishing posts for this website and developing new things for the Portable Antiquities Scheme website that I develop and manage. The actual ‘Day’ for me was an interesting affair that started the night before working till midnight with some tinkering with the site to iron out some bugs that has pre-released over 20 posts from RCAHMS (these were fantastic) and then rescheduling them following the discovery of the problem (an incompatible plugin) and then started again at around 5am when my son woke me up:

Then a fast cycle into work at the British Museum. Little glitches were identified in some of the plugins and these were fixed, probably without anyone noticing and the workflow for getting posts seems to work  well. Throughout the day load and activity on the server was monitored, we had no real problems and Tom Goskar asked for a cache to be enabled in case we had a surge in activity.

Whilst not editing and publishing posts via the scheduling feature, I was working on my current development work, which is an extension of the LAWDI summer school programme I participated in. I’m modelling Portable Antiquities Scheme (PAS) data to the CIDOC-CRM mappings that the British Museum have used to allow our data to be harmonised by the ResearchSpace project. I’ve been linking PAS records to the URIs that exist in the BM system, in the Ordnance Survey data endpoint, to Geonames, to Nomisma, to the thesauri exposed via Seneschal (see this post by Michael Charno at the ADS for more insight into what they are doing there). I think I’m getting there and you can see this N3 view of one of the records linked here (it might take a while to load as this is an external service), if you see problems with what I am doing tell me as I’m tinkering in the dark, some URIs are missing their identifier off the end of the URI string as I haven’t updated the search index on that server – for example OS ones. Once I get this working properly, we’ll have over 560,000 records in RDF format, who knows what people might do with the data – serendipity is king as my good friend Vuk is wont to say.

Enjoying the outputs

Running through the posts, many caught my eye. The content was fantastic (over 300 posts), the images (over 1,100) amazing and some of the commentary coming in (not the pingbacks) was insightful. For me, some stand out posts:

There’s too many to mention, and the LAARC ones were excellent, INRAP’s contributions ace. Every entry is superb in its own right and Janet Davis summed up the event succintly:

Back Channel

As usual, we tried as a collective to maintain a healthy presence or back channel (you can read more on this idea in this pdf by Ross, Terras, Warwick and Welsh) on social media using two platforms – Facebook and Twitter. In my eyes, the Twitter platform has been more productive (even though we gained fans/likes on Facebook). It was easy to measure whether links were being clicked on as I set up a plugin that automatically tweeted the majority of posts (except for when we exceeded the rate limit for daily photos being posted – I didn’t even know this was limited) and shortened them to a url. Over 5,500 tweets (inc retweets) were sent using the #dayofarch hashtag – to put this into perspective, the British Museum #pompeiilive archive that I collected showed 18,000 tweets relating to their cinema extravaganza. These tweets were collected using Martin Hawksey’s  Tags Version 5 tool which is easy to set up and the only tricky bit is setting up the authorisation with Twitter, and then the conversation could be analysed. For example we could see how many people used the hashtag in their output (696) and who the top tweeters were and how many interactions or @ were made to them using the hashtag:

Top Tweeters Volume of tweets @’s % RT
dayofarch 619 4917
AdamCorsini 132 180 17%
lornarichardson 124 209 31%
portableant 122 164 32%
rcahms 121 170 13%
m_law 83 90 33%
tharrosinfo 81 3 81%
JaimeAlmansa 78 32 23%
TRArchaeology 75 8 67%
TinctureOfMuse 69 11 61%
VitaEmilia 67 48 10%

And then we could see what the network graph looked like (this one is with mentions clicked in the bottom right corner):

TAGSExplorer  Interactive archive of twitter conversations from a Google Spreadsheet for  dayofarch

And what the timeline looked like for posting frequency:

TAGS Searchable Twitter Archive

I’ll be doing some more analysis of the Twitter archive using the programming language R shortly.

Running the project

The ‘Day’ as a concept has definitely been fun to help co-organise with a fantastic team of people over the life time of the project; for 2011-12 iterations we comprised the collective of Lorna, Matt, Jess, Stu, Tom and Andrew and myself and then this year we changed slightly with the inclusion of Jaime (who made great efforts to branch out into multi-lingual contributions), and Monty Dobson. We lost Jess, who has just got married to Leif (congrats you two) and Stu along the way. The team has functioned really well. If you’re interested in how we’ve managed to keep this show on the road, a combination of tools have been used:

  • Basecamp
  • Google+ hangouts
  • Skype
  • Twitter
  • Gmail
  • Very infrequent vis-a-vis interactions as we’re a team divided by oceans

The site itself is quite straightforward. We run on:

  • a wordpress installation (even though if you look at the HTML code under the hood, you think spaghetti code) using the latest version (at all times!)
  • search is provided by the solr for wordpress plugin (which is pretty powerful and allows the faceted search)
  • the theme (overseen by Tom Goskar) is from WooThemes and is the Canvas version
  • we use OpenCalais for generating tag suggestions for post (by analysing what you have written in your contribution)
  • for posts submitted by email, we use the Postie plugin (this is superb, but you do need an account first before your post will be accepted.)
  • tweets, vimeo and youtube video links were easily converted just by placing the url in the text of a post (no need for embed)
  • Akismet stops spam comments coming through (there’s so much spam out there.)
  • A linked data view of the posts can be generated via the wp-linked-data plugin

If you’ve got any questions about the technical side, do email me (I’m easy to find on Google).

Reflection – my opinion (not the collective)

But, have we made a major impact? Reflecting on the ‘Day’ as a project, yes, we have made an impact in some ways. Readership has not been massive, the Google Analytics figures show interaction magnitude of 1000s rather then 10s of 1000s (5,818 visitors on the day). However, the people that have taken part have made a concious effort to participate and I hope that everyone that has participated has enjoyed it? Myself, I’ve been flamed on blogs for my contribution to running the site and my integrity questioned, and the author of those offered nothing to the site about his archaeological day or any positivity at all. You’ll know where to find them if you’re associated with archaeology and metal detecting debates.

I’m disappointed that more of my colleagues from the Portable Antiquities Scheme and the British Museum haven’t contributed to this project (thank you to Julie Spencer, Jonathan Taylor,  Ian Richardson and  Peter Reavill for taking the time out of your working day to join in), seeing as both of these organisations were supporters of the project. I believe that this is a good project and hope that it continues for a few more years at least. The resource created, by you, the contributor, is amazing. An insight into the world of archaeology that isn’t available anywhere else in a searchable, discoverable format. It is even available as linked data.