Some of us are lucky enough to be attending the VALA Conference. For everyone else, you can follow the Twitter stream, and I’ll be writing up notes from each day here.
We began Tuesday with a keynote presentation from Christine Borgman called “Big data, little data, no data”. Aimed primarily at academic librarians, she made the point that data and academic publications are not the same thing. Data is evidence for arguments which are articulated in publications. Publications have clear copyright owners and are written for impact, whereas data ownership is often unclear and the representation (and therefore understanding) of data changes depending on the method used to represent it. Researchers are often reluctant to ‘open’ their data because they want to explain it before it is used by someone else. This tacit knowledge (“don’t use it for x because it won’t really work that way, but it would be great for y”) is difficult to capture and use in a scalable way.
In Session 5 Simon Cootes from the State Library of NSW told us about their Wikipedia project where they created pages about over 80 digitised NSW newspaper titles in Wikipedia. This promoted their collections, developed staff skills and led to a 20-30% increase in traffic to Trove and the SLNSW website from Wikipedia. Wikipedia is the sixth most visited site on the world wide web and the second most popular information source for students – we need to be where the people are.
All the staff involved volunteered and they came from all across the organisation. Most of the time spent on this project was actually the research they needed to do, rather than writing the Wikipedia articles. The project involved some local libraries updating records with more detail and the wider public have also contributed. One thing some librarians may be uncomfortable with is that once the article is up on Wikipedia it doesn’t belong to you any more, it belongs to the general public who can alter it. The project has, however, seen a great deal of improvements and corrections to information from members of local communities who have more local knowledge about some of these historic publications.
This is a great model we can learn from – the SLNSW took a specific topic (historic newspapers) and improved the common resource of Wikipedia and the visibility of their resources at the same time.
Next up we heard from Edmund Balnaves who is not a character from a Victorian era novel but actually a software engineer from Prosentient Systems. He told us about a project they did with the Parliamentary Library of NSW to harvest web and email content relating to the NSW Parliament – things like MPs’ media releases and news articles about Parliamentary business. The system needed to provide a simple workflow allowing staff to make a quick ‘accept/reject’ decision, be integrated with their existing systems and be low cost.
They built a bespoke solution on web standards and mostly open source software, integrating data harvested via RSS, OAI/PMH, JSON, Schema.org, XPath, search APIs and email APIs. It was complicated! Not everything could be automated – there’s no substitute for the human eye sometimes.
After lunch we heard Dianne Velasquez from University of South Australia talk about some research she did into usability of public library websites. Horrifyingly, they discovered that a huge number of sites are really terrible. The looked at over 120 websites, mostly in SA, NSW and QLD. The majority of websites were missing basic data like date last updated, search box, feedback mechanisms, a library image and Web 2.0 tools.
Next up was Tony Iezzi from Vision Australia. He gave a really interesting presentation on making websites accessible to vision-impaired people. The key points were that you really need to think about your page structure and metadata, because semantic markup is what screen-readers use to help the user navigate the page. When website structure is changed it really confuses people using screen readers because they have to re-learn the navigation.
Last in this session we heard from staff from Auckland Library of Technology. They told us how they designed “a library website that’s not for librarians”. It was a really useful explanation that essentially boiled down to ‘test the usability of your website with the people who will actually be using it’. Revolutionary, huh? They aimed to remove as much as possible from the site to make it clear and easy to understand. ‘Heat map’ analysis showed that of the 30+ links on the front page only about 5 were used with any regularity. Their new site is at http://www.library.aut.ac.nz – as you can see they did a great job.
We ended the day with a keynote from Johan Bollen. Johan told us about some research he’s been doing tracking the aggregate emotional state or sentiment of Twitter. He and his students have worked out a way to predict the stock market using the ‘big data’ of Twitter’s emotional state!
All the Conference Proceedings (including video of each session, and the paper where relevant) will be available to the general public in 3 months.