Thursday, 18 November 2010

Addressing History Launch - report

On Wednesday I was very kindly invited to attend the new AddressingHistory website (http://addressinghistory.edina.ac.uk/) launch at the National Library of Scotland. I've already blogged with a short video of the event's Nicola Osborne yesterday, but now time for some more detail...

The event started at 2.00pm and included several presentations, mainly linked in some way to the project, though not exclusively. If there were new buzzwords that I learned from the day, they were "crowd sourcing" and "georeferencing". Various definitions for crowd sourcing include 'delegating a task to a large group', or 'a process in which readers submit individual reports that are collected into a larger dataset for use in reporting a story'. Georeferencing can be defined as 'to design something's existence in physical space, relative to a standard reference grid'. I thought I'd briefly list those, as everyone was merrily using the terms in the way that I never do!

Addressing History is a site that uses historic Edinburgh maps which have been matched up or georeferenced to a modern Google map. Entries from three directories - 1784/85, 1865 and 1905/06 - have been pegged to the relevant locations on these maps. The combination provides for a potentially powerful search tool - if I want to search for all photographers in 1865 and then compare to the same search in 1905/06, I can perhaps observe a shift in the location of the photographic community in those eras, for example. The crowd sourcing comes into play in the way that the general public is now being asked to interact with the site, to help improve the work of the OCR (optical character recognition) scanning and more - in the same way that the Australian newspaper site from the National Library of Australia allows for users to go in and make corrections to its OCR scans which are displayed in parallel with the digitised images.

The event started with Professor Robert Morris from Edinburgh University, perhaps an odd choice for the first talk in that he defined many of the flaws of the AddressingHistory site before the AddressingHistory folk could actually define what the site was! However, there was a serious point to his talk. The new site is still very much a beta project with much development still to go through, and his talk raised many of the challenges that the project has had to contend with and which it still has to deal with - not least of which the use of OCR which at times has experienced a 40% error rate. There are also several restrictions and quirks of the directories themselves that can create some issues - the occupation search field, for example, is not much use if the women listed in a directory are not listed with an occupation, and the fact that it is difficult to compare like with like between directories, as the directories themselves changed so much from one version to the next. The professor's summary of the site was that it was "a ship ready for launch but still not quite ready for sea". An interesting presentation, but I'll admit to not being quite sure whether the good professor's glass was half-full or half-empty!

Stuart McDonald and Nicola Osborne then took the floor- both work for EDINA and are heavily involved with the project. Stuart explained that the AH project had been funded by JISC from April-September 2010, and partnered by the the National Library of Scotland, which sourced the Post Office Directories(or 'PODs' as they were called throughout). Nicola mentioned that the first phase of the project was to physically link the data with the maps, and that the next stage now was to launch the site and to ask the public to help "get the data up to scratch" - the site itself went online officially at 3pm during their chat. It was announced that EDINA will continue to support the project for at least the next year, and that models were currently being investigated for the longer term future of the site, possibly through donations. But now is really the evangelising phase - get the word out, and show the potential of the site. Nicola also mentioned that she is particularly interested in further guest posts for the AH site's blog.

Professor Richard Rodger of Edinburgh University then took the floor to talk about the Visualising Urban Geography project at http://geo.nls.uk/urbhist. This is like a much bigger version of Addressing History, in that it is trying to link data from various sources to contemporary maps. An extraordinary project, but rather than have me attempt to define it, visit http://geo.nls.uk/urbhist/examples.html for several examples of projects being worked on!

There then followed a presentation by Helen Chisholm on the Statistical Accounts of Scotland, as hosted online by EDINA. I've used the accounts a lot, but not the subscriber's version of the site. Very interesting additional features, but I will post on this on my Walking in Eternity blog (http://walkingineternity.blogspot.com/) in the next few days. Suffice to say that there are some developments on this front, so stay tuned.

Then followed a master session on georeferencing from the NLS's Chris Fleet. Chris explained how sites such as Google carve up the world using a programme called Map Tiler, and how this forms the backbone of several current NLS mapping projects. In addition to the OS maps online at the NLS website, he also explained that the library has a map collection called OS Mastermap available at the buidling but not online, which holds the most detailed OS maps in Britain. Other projects using georeferenced map dat from the NLS include the gazetteer site at www.scottish-places.info, Heritage Paths at www.walkingthroughtime.co.uk, the Portable Antiquities Scheme site at http://finds.org.uk/, the Society for All British and Irish Road Enthusiasts site at www.sabre-roads.org.uk (the history of roads - amazing site!), ScotlandsPlaces at www.scotlandsplaces.gov.uk, the Stravaiging Around Scotland site at www.stravaiging.com and many more, a few of which I had never heard of before but which I will be plundering in due course! Future mapping additions to be added to the NLS's own maps website will include 6000 sheets from the OS 6 inch and 25 inch maps from the 1890s-1940s, as well as many Bartholomew's maps.

Native Uisteach Kenny Beaton then gave a talk on the Tobar an Dualchais / A Kist o' Riches project, accessible online at either www.tobarandualchais.co.uk or www.kistoriches.co.uk. This is a £3 million audio digitisation project which ran from 2006-2010 and which has digitised 12,000 hours of Gaelic and Scots recordings (the equivalent of 3 1/2 years of listening). The material includes 10,000 hours of recordings from the School of Scottish Studies, the National Trust's Canna Collection (400 hours of recordings by John Lorne Campbell, inclusing material in Nova Scotia), BBC Scotland output from 1950-2005, and more. Each recording was digitised at 4 times the quality of a CD for archival purposes, requiring some 12 terabytes of storage. Not all is available online, but you will find 50,000 tracks available, and free of charge. Although the site does not officially launch until December 9th, it is up now - well worth a visit.

The most enjoyable talk of all was by the National Library's Ines Mayfarth. Ines gave a talk on her work to digitise the 750 Scottish directories from the 1770s to 1911 for the Internet Archive, covering Edinburgh, the Lothians, Glasgow, Stirling, Perth and Kinross, Dundee, Angus, Aberdeenshire, Invernessshire, Moray, Renfrewshire, Bute, South Ayrshire, Inverclyde and Dumfries. As well as material from the NLS, various partner organisations are helping to plug gaps in the NLS coverage, though Ines did mention that the Mitchell in Glasgow was still to be convinced! Hopefully the Mitchell will get fully on board, because frankly, the state of some of the directories at the Mitchell leaves a lot to be desired (some of their directories are literally falling to bits), and digitisation should really be a priority. Ines and her team scan 24 images per minute, and she describes the OCR accuracy of the scans to range between 50% and 90%. To help with searching, the NLS is creating a new index of the 1st three letters for every surname on each page within each directory, and the first books will be made available on the NLS website by Christmas, in the family history pages area, and by Autumn of next year a dedicated site will be launched for all the directories.

It was a great event centred around Addressing History, and an exciting glimpse of what the NLS sees as the way ahead with some of its key projects. Thanks to the team for the invite, and best of luck with the project's furture development, because it has some seriously useful potential, not just for academia but for the humble family historian also!

A report of the event by Peter Munro of Borders FHS is also available at http://blog.bordersfhs.org.uk/2010/11/addressinghistory-launch-event.html
Chris
www.ScotlandsGreatestStory.co.uk
Professional genealogical problem solving and research
http://twitter.com/ChrisMPaton
Researching Scottish Family History (New book)

No comments: