Firstly, thanks to those Campers who came to my roughly planned session and who gave such great input.

Although the purpose of the session was to talk about this issue across the GLAM sector as a whole, we used Archives New Zealand’s electronic finding aid Archway as a “test case”.

Following is a list of possible solutions the group came up with, ranging from very manual to more automated:

• Capture contact details of researchers and their interests – contact them when new material becomes available
• Set up webpage to publicise new material
• Create an RSS feed of newly added items
• Something similar to the trove news bot – easy to use, it checks for messages from Twitter to create queries in Trove’s newspaper database, tweeting the result – still requires a pull by researchers
• Provide ability for researchers to save search criteria
• Provide functionality to highlight/identify those search results that have been presented to the researcher previously; ability for researcher to filter out these out of the result set

We also discussed other technologies currently in use, or provided, by various institutions:

• British Library’s Mechanical Curator – undirected, haphazard, unplanned publishing of content
• Digital NZ’s custom search builder – a tool that allows anyone to create a mini search engine across Digital NZ’s aggregated digital content, also to create an embeddable widget to share the content
• Academia.edu – allows registered users to add research interests to their profile, and delivers content (research papers, etc.) that has been shared by other registered users and tagged with the chosen research interest phrase.
• Google Scholar and other library search tools – a post THATCamp investigation of Google Scholar revealed that the tool uses “robots” or “crawlers” to fetch files from websites for inclusion in the search results.  This is the type of thing I was wondering if researchers could create for themselves.

Discussion was had about the need for not only appropriate tags against each item of material so it can be categorised/classified, but also the need for metadata that will enable capture of when records are added / updated so the researcher can filter out those items they may have already seen.

The group also identified some possible funding/resourcing options:

• Create partnerships with open source development organisations
• Collaborate with National Library of New Zealand (this option relates specifically to the Archives NZ case)
• Apply to Internet NZ for funding
• Create a research question for an information studies masters student

Many institutions have a backlog of “items” in their holdings that are not currently discoverable/accessible in their online finding aid(s). Most institutions are actively working on rectifying this, as funding, staffing, etc., allow.

Currently if researchers want to know if more items relating to something they are interested in (e.g., person, place, time period) can now be found via an institution’s finding aid, the researcher has to decide how often to got back to the institution’s finding aid and re-submit their search criteria.

What if, instead, the researcher could submit the search string once and include in that search the request that he/she is automatically advised if anything new results from that search string are returned.  A bit like TradeMe searches. Or perhaps the Digital NZ search builder (digitalnz.org.nz/about/custom-search-builder).

What functionality would researches like? Are there existing add-ons that institutions could use? Or are there apps (I’m thinking ‘bots) that researchers could use to achieve this?

Background

Data collection from human subjects is always a challenge. Now that everyone and their dog is on the internet, it has become possible to collect data online. Some researchers have investigated the feasibility of doing so and have concluded that it is possible. These data can also be analysed to understand human learning, cognition, psychology, and possibly other topics of interest.

Most of the researchers seem to be paying participants small sums in exchange for the data they produce. But another alternative exists when you consider what is going on with all the data that users generate on interactive websites. Many companies mine it and sell it to advertisers, market researchers, and the like. Others use it to improve the user experience or to evaluate changes to the code base.

Proposed discussion

I’m interested in possible ways of combining the provision of useful services in exchange for collecting (mostly) anonymised data which can then be used for research. Twitter, for example, has created one of the largest corpora in history of speech-like text which everyone from computer scientists to linguists to political scientists are analysing.

Other examples include Coursera and Khan Academy, both of which collect data on human learning in exchange for a free education. Other sites, such as Human Benchmark, don’t even really offer a service, and yet manage to collect impressive data sets.

So, what I propose is a discussion about

• what types of human data are interesting, but difficult to collect
• what kinds of services or formats could be used to entice people to produce that data
• what are some reasons why these types of services have succeeded/failed in the past
• what existing platforms/projects could be leveraged to facilitate data collection and service provision
• etc.

Qualifications

I have some limited experience collecting data through websites which have yielded interesting insights into vocabulary acquisition and rent pricing (yes, they are completely unrelated). I’m keen on pursuing this concept further to study the development of reading proficiency and speed in a second/foreign language.