Here at SHU there’s a couple of PhD researcher competitions on at the moment as part of the forthcoming Doctoral Showcase series. There’s the ‘Three Minute Thesis’ heats and local final, but the one that attracted my interest was the ‘SHU Doctoral Research Image Competition 2018.’ I’ve been producing visualisations throughout my study and I had in mind one I wanted to produce, but hadn’t because I knew it would suck up time. The competition provided the final impetus and although I suspect from the information and instructions, the organisers are expecting photographic images, I thought I’d have a shot at pushing the boundaries.
We welcome attention-grabbing images to intrigue, inform or excite a lay/non-specialist research audience about your research. Images may be arresting, beautiful, moving or even amusing but they must relate to your doctoral research project.
Entrants are also allowed 150 words of accompanying text; here are mine:
The flâneur of 19th Century Paris was an observer and chronicler of city life. In exploring the bold claims some teachers make that ‘Twitter is the best PD ever!’, I called on the spirit of the flâneur to guide my ethnographic approach.
One of several methods I employed in the study was participant observation; this image is formed from tweets collected during that process. Each of the districts or ‘quartiers’ contains tweets on one of the emerging themes, each typified by a magnified example.
Since flânerie inspired my approach to observation, analysis of the data, and presentation of the findings, I sought an image which spoke to that activity. Although somewhat playful, creating this image, and other visualisations during the study, was more than simple representation. On each occasion I found the attention to compositional detail which was demanded also yielded additional analytical insights.
Having written about flânerie earlier, it’s developed into more of a ‘thing’ within the thesis and helps to describe and explain my ethnographic approach, together with the analysis and presentation of the data. As with other visualisations which have appeared on here, I found producing it the most valuable part of the process. As you design and construct the vis, you’re constantly thinking about what goes where and why, and what will it say to its audience? Once released, it will take on a life of its own for those who view it, each seeing it and interpreting it in its own way. My own view will be only one amongst many.
I guess the first thing to say is that it only produces a partial view of the data, focusing as it does, solely on tweets. The inspiration came from trying to imagine how I could convey a sense of what a Twitter flâneur or flâneuse might have encountered on their travels. Some of the tweets appeared as a result of wandering through my timeline, some came from searches performed by me and @PLDbot, and others came from hashtag chats. There were several thousand in all, but I wanted to consider different ways of exploring them rather than by coding and categorising them into themes. As you can see from the image, I didn’t entirely succeed in that, but perhaps the route by which I came to those ideas was slightly different. As flâneur, I wandered through the tweets in the same way that a metropolitan flâneur might wander the streets, noting interesting features and activity. Similar observations might occur later on the journey or on a different walk and gradually within one’s notes and memos, they take on greater significance. For me, that’s a different knowledge generation process to iterative coding and theming, one which feels more in keeping with actor-network theory in which knowledge is assembled as connections are made. Any one tweet might take me in a particular direction, up a different boulevard if you will, which then brings me to another tweet and another and ever so slowly a fragile picture begins to emerge. As more tweets are added, it is like more paint being applied to the canvas, and the picture takes on a more durable form.
To translate that emerging knowledge into a visualisation, I was keen to try to keep some essence of what the sense-making process had involved, melded with where my thinking currently resides. There were two sources of inspiration to whom I am indebted, and it was rather appropriately through Twitter that I came to them both. I have Deborah Netolicky, the édu flâneuse, to thank for bringing me to flânerie in the first place; having read her blog and been impressed by her educational observations and insights, I also became fascinated by what flânerie involved. I am so grateful to Martin O’Leary for his @metropologeny Twitter account/bot which automagically generates hourly blank maps of ‘Cities, towns, villages, hamlets, conurbations, burghs, municipalities, metropoles.’ It is fascinating and you should check it (and his other projects) out.
So to the vis. The nine themes exemplified by nine larger tweets only provide a partial picture of the findings. There’s more to come, but that’s for the thesis. All but one of the authors of those tweets gave their permission for me to use them in this way, so I’m grateful to them too. In one case, I didn’t get a reply, so took the ethical decision to anonymise that tweet. Given the time constraint I was under, I didn’t have the time to seek permission from the 100+ authors of the tiny tweets, so these are all presented at lower resolution, hopefully to anonymise their authors too … though you might be able to spot one of your own tweets of course. In the interests of full disclosure, I have to admit that, although all the tweets have come from the data I collected, I ran out of time matching each of the tiny tweets to the relevant Quartier. The vis is more about the bigger picture though, rather than specific detail. I wouldn’t want for example for people to think that because the ‘Discursive Quartier’ is bigger than the ‘Pyjamas Quartier,’ and it has more tweets, it is somehow more important within the findings. Let’s not forget that I’m not aiming for a realist account, though acknowledge that someone could take these ideas in that direction. The consequence of that restriction however proved somewhat serendipitous …
I was somewhat troubled by clustering the tweets into Quartiers in the first place. Some viewers might take it that whilst out wandering, whether the streets or the data, the flâneuse encounters all the tweets in neat bunches or districts. In the same way that the Cultural Industries Quarter here in Sheffield, or Jewellery Quarter in Birmingham, also have other other buildings and activities taking place within their environs, so too the Quartiers within the image. Owing to time constraints, I had to start siting tweets where they ought not to be expected, which perverselyseemed to make more sense. The Quartiers, although presented as homogenous with clearly defined boundaries, are only notional areas at best. My attempts at making sense of them are of course consequently also notional. Another researcher with the same tweets would likely take a completely different approach and produce different knowledge. I’m comfortable with that, although those with other epistemologies may not be of course.
Other features of the vis began to jump out as it began to take form. Tweets were positioned largely according to their shape, rather than as a result of their content, so when tweets came together that might not otherwise have done, did that yield new insights? I couldn’t help wondering here if this aspect of an analysis might be lost when moving from shifting around physical lumps of data, to having a CAQDAS package doing the manipulation and filtering.
On the underlying base map, there were areas of space in the city, so I tried to mirror that in the tweet map. To me, this spoke about gaps in the data and what might be left unsaid; things a researcher might expect to find, but were not there. I also found it frustrating at first, trying to fit rectangular tweets into city blocks which were rarely regular and often had curves; it offended the (admittedly naive!) designer in me. It would have been relatively easy to skew, stretch and transform each tweet to mirror more closely the borders of the block into which it had been placed. This immediately pricked my ethical conscience; that would have felt like I was altering the information that participants had provided. Although researchers do that all the time as data is interpreted, distorting the tweets, reshaping them, seemed more of an intrusion – I didn’t like the way you said this so I’m going to change it so it better fits my interpretation. The base map also has buildings of different sizes, so I’d resized tweets to reflect that and even this distortion bothered me that I might somehow be implying that some tweets ought to be more prominent than others.
Looking at the vis in toto, I was struck by how much those tweets with embedded images stuck out. It will come as no surprise that tweets containing images do grab attention more, at least according to those involved in online marketing. It is arguably much quicker to instantly, if not always correctly, get a sense of something from an image than a section of text. Teachers do share photos and graphics of the resources they share, the activity in their classrooms, the displays they use, and much more. Twitter has made adding an image so much easier over the years, and the code which presents a tweet in the timeline also includes a thumbnail or larger image clip. Whether you’re marketing a product or sharing a teaching resource, including an image does seem to make a difference. Admittedly I didn’t go looking for that in my data, but my reaction to the image tweets in this vis might provide a pointer worthy of further exploration – future research?
Overarching all these observations though, I guess I ought to ask myself what I learned about teachers’ professional learning practices. As a flâneur wandering through this city of tweets, I’d clearly find an eclectic mix of different practices all closely intertwined, just by turning from one street into the next avenue. There are things I might see as flâneur that I wouldn’t as a commuter rushing from tube station to place of work. Just as some city districts are no-go areas for some people, perhaps there are strands within these finding that wouldn’t have emerged by other means. Should hygge be considered an element within professional learning practice, or should I have consigned it to an outlying district? The suburbs maybe?
I’m now considering ways II might push the vis further, given more time, but like all the other visualisations I’ve attempted, perhaps it served its purpose. It made me think … and every time I look at it, I’ll think some more.