A quick ponder into visitor generated content and value classification.
Twitter has introduced new metadata for tweets; with the objective of helping developers filter out the most “valuable” tweets. This immediately got me thinking about visitor generated content (VGC) in museums. My PhD is grappling with the idea of impact and how you can go about measuring impact of VGC on museum experience. Over the past couple of years working on QRator and the Social Interpretation project, it has become clear that VGC, impact and value are notoriously difficult to define, interpret and well basically study.
In essence, Twitter is going to be introducing new metadata for Tweets so that you will receive tweets tagged up with value levels; initially just no value, low and medium. No High value tweets just yet. The aim is to make it easier for developers to surface what is arguably the better and more interesting content from otherwise noisy or high volume tweet streams.
We had a similar problem with the Social Interpretation VGC in particular, a high volume of visitor comments, and no clear way of moderating, categorising or “valuing” the better quality visitor comments. As with most high volume unstructured data, finding and highlighting the signal out of the noise can represent a significant challenge.
The problem is that “value” is highly subjective and varies on the context within which it is being consumed. One visitor’s value is not the same as the next. Nor is it likely to match what the museum defines as adding value. The SI team at IWM experimented with gardening comments, but we didn’t come up with a criterion to work from, so it was up to the moderator at the time to decide. At the Grant Museum with QRator we are trying to come up with criteria to look at the visitor answers to the current questions, but this is after the point of visitor contribution, and is very much based on the museum’s perceived value of the response. Is value something you add in the post moderation stage? Who’s value? The visitors or the museums?
So, is Twitter’s new value algorithm something that can be used by museums to classify VGC?
If I’m honest, no I don’t think it is. Is it really possible to create an algorithm that can classify value of comments? Surely value is judged by the reader? Can an automated system really evaluate subjective factors and identify the most valuable conversations for each individual? Doubtful.
But I will be watching how Twitter deals with concepts of value of tweets with interest.
UCL runs a series of public events under the title of Brain Food, which I think is a brilliant title. Part of the Brian food events are the Lunch Hour Lectures. The lectures, which have been running since the 1940s and rated as one of the 10 best ways to spend your lunch break in London. The main aim of the LHL’s are to provide an opportunity for anyone to sample the exceptional research work taking place at the university, in bite size chunks.
The 17th March saw Claire Warwick’s Lunch Hour Lecture to mark 5th anniversary of Twitter on 21 March. I had the privilege of writing the LHL brain food blog post about the lecture. It will be going up on The LHL page I think. But I’m not sure when or how. But here is a bit of a sneak peak of part of it.
LUNCH HOUR LECTURE: Great 2 meet u IRL Twitter and digital identity
The topic focused on Twitter and digital identity. Asking questions about whether or not Twitter is an ephemeral technology, consisting of mundane, pointless tweets about people’s personal lives, normally revolving around what people had for lunch? Or can a study of its use help us to understand how we express our identities on and offline?
Claire Warwick started the LHL by stating that each of her slides will use Twitter conventions and language, all the slides were also auto tweeted so there could be a direct online discussion as well as followed by those in the lecture itself. Traditionally lectures have a single focus of attention. I think Claire showcased how Twitter, positioned alongside the official lecture, can provide another dimension of communication which can extend beyond the lecture room to engage with a much larger audience. Changing the dynamics of the lecture room from a one to many transmission to a many to many interaction. The LHL’s are live streamed so this also adds another type of audience into the mix. It is obviously incredibly important to include that audience in with the discussions arising from the lectures, just because they aren’t in the physical space does not mean that they are in anyway less engaged and interested in the content. Twitter provides that facility.
Claire’s approach to the lecture was lighthearted, again in with the way a lot of academics, particularly at UCL’s Centre for Digital Humanities use Twitter. There is a serious point to this though, Twitter conventions are an interesting challenge when taken out of context and put into a physical space, but Claire shows it can be done, raising questions about the fact that there are different languages within digital media. Tipping these language conventions on its head by deliberately changing the way you register with a language and a tone of voice. Claire also stated that Tweets are More permanent than speech which is something individuals should be aware of.
Claire then went on to discuss Digital Identity Asking: Who are we online really?
Is the UCLDH website, or the UCL website part of our digital identity? The answer is most definitely yes, Corporate Digital Identity is becoming more and more important. But the purpose of this lecture was to talk about individual digital identity. This is not a new idea, it has grown from the concepts of cyborgs and avatars and talking to the machine to brilliant manifestations of virtual communities to tweets & social media.
But what does that do to our identity? Are we playing a role or being ourselves?
The idea that the avatar an individual chooses actually says a lot about you; some of the options discussed were:
Yesterday I went to my first CASA seminar, and it was great! Well the second half was, the first half involved a lot of equations about Thermodynamics and I didn’t have a clue what was going on, and I chose an unfortunate seat near the front, I was slightly terrified that he might ask the group to solve said equations. I have trouble adding up numbers let alone Greek letters. I’m sure it was brilliant, but unless you have a firm basis in maths or thermodynamics I don’t think you would have stood a chance. Then came the good stuff, the stuff I understand, and the stuff that makes me happy: Social Media, visualisations, maps (I am a closet geospatial nerd who has no geospatial abilities- I’m like superman on kryptonite) animal logos and to my happy surprise a bit about museums.
Steve Gray and Fabian Neuhaus provided an overview of the tools in CASA’s crowd sourcing toolkit; SurveyMapper, Tweet-o-Meter and the Twitter Collection tools. There has been a massive explosion of handheld mobile devices with GPS as well as a move to crowdsourcing info this has produced a heck of a lot of online geospatial data. Add newly released public sector data and you get yourself an exciting situation where people can take that data and turn it into something more interesting. CASA work on integrating tools for unlocking, exploiting, understanding and sharing new data sets and to also enable users have a go at mapping and spatial analysis.
Firstly Steve talked about Survey Mapper – a real time geographic survey tool. What I like about survey mapper is that it doesn’t take itself too seriously. It knows it’s doing clever stuff behind the scenes, but presents a friendly giraffe – you can’t not love a giraffe – with an easy to use interface. One of the survey’s Steve discussed was the BBC’s Look East Broadband speed survey which produced a lot of responses over 6500 in a day I think (I might be wrong on that one).
Tweet-o-meter – This is genius and really beautiful too.
Tweet-o-Meter is powered by CASA, as part of the NeISS project. Created by Steven Gray
Tweet-o-meter harvests geospatial data from Twitter with the aim of creating a series of new city maps based on Twitter data. Data is collected from tweets sent via a mobile device that includes the location at the time of sending the tweet. Via a radius of 30km around different cities, for example the number of Tweets have been collated to create New City Landscape Maps of London, New York and Paris.
Created by Urban Tick
I think this is a beautiful analogy for twitter activity where contours correspond to the density of tweets, mountains rise in active twitter locations and cliffs drop down in to valleys of tweet deserts.
UrbanTick has the full set of the different new city landscapes, all available in Google Maps viewer (I think)- head over to take a look at the gorgeousness.
Steve and Fabian discussed that there are now 60 cities around the world that have their tweets monitored over the period of one week. Amsterdam is a top tweeter with over 50% geolocated tweets. Whereas London which is still a really active city send on average send about 10% geolocated tweets. Visualisations clearly showed that different cities seem to be more active in the morning and others in the evening, producing some lovely looking kidney shapes. The data also shows that different days of the week are more conducive to tweeting, for example Monday and Tuesday are generally less active than the rest of the week.
Data was also collected during the early days of the Egyptian revolution in Cairo. It was really interesting to see how the protests and internet blackout affected twitter activity. For example when the big internet switch was flicked back on the data shows an immediate rise in geolocated tweets.
And then came something that I got really excited about and something that I could really use in my PhD… Andy, Steve, Fabian … if you’re reading this, can you show me how to do it? Pretty please!? I will buy you cake. Lots of it.
Tweeting art – Most museums are now using Twitter and CASA have taken that information and turned it into really awesome spider like explosions of communication network visualisations. Showing how different museums (the examples given included Tate and MoMA) link in to the wider twitter network and also how they link to each other so in essence how the institutions interact with other users and how this connects them into an entangled social network. For example Tate and MoMa tweet to roughly the same followers but don’t really tweet to each other. I think this is fascinating, particularly if it can show if museums are only using Twitter as a broadcast medium – pushing marketing out, or whether they are creating a engaging discussions and digital experiences with their followers.
It was fascinating to be part of the seminar not only where people were talking about Twitter in a active exciting research context and sensible manner, and where questions from the audience were serious, and probing and engaged in the topic. Rather than asking ridiculous questions from anti social Media people where Twitter is a waste of time, full of pointless babble which makes yooths mediocre there was brilliant questions about how do you model for uncertainty, what proportion of users geotweet? Does this skew the data?what about frequency and text mining to find out more about context. It was brilliant. I was engaged. And it reminded me that I am not out of my depth in this whole digital humanities thang, I do know what I am talking about, and this is a growing research field doing so much cool stuff, I am an alt academic and proud.
Our ( Melissa Terras, Claire Warwick, Anne Welsh and me) Paper “Enabled backchannel: conference Twitter use by digital humanists” has just been published by the Journal of Documentation! And I’m first Author!! Excited Muchly! It’s my first proper paper. I am chuffed.
As I have mentioned previously I attended the Autopsies group studay day Yesterday’s objects: the death and afterlife of every day things. Myself and Ernesto tweeted our little hearts out during the event and the Autopsies group liked it so much they asked our permission to display our tweets, which have been compiled and can now be seen on the Autopsies group blog (warning we tweet a lot – its quite long). Everyone should check out the Autopsies group they are doing some pretty cool stuff.
Claire Bailey-Ross 