Would be possible by introducing one more layer in the content hierarchy of what Gource is fed. Then it would be: “Community → Category → Subcategory → Topic → Replies”. I think @alberto might have misunderstood what you want to show?
Yes, that’s what I mean.
1 Like
5 Likes
No, no, that would be easy. The issue is that it is not so intuitive to see if people are moving around as opposed to staying in their home forum.
is there any way you could replace the monopoly pawns with cats? Asking for a friend.
4 Likes
There’s this interesting effect now of seeing Gource at larger scales, where people become too small to see and observing the day-to-day stuff happening becomes meaningless. But visually it’s still easy to see patterns … which human visual cortex is made for … and to observe the “order” that is emerging from the far-too-many detailed interactions.
Haven’t seen Gource being used at this scale anywhere so far (and I’m amazed the software did hold up so well doing that work).
The animation came out much better than I thought. Great work @alberto 
2 Likes
I like it a lot! Was it fun?
I think it could help to:
- Include millenia at year at the top (12-05-2020)
- Use labels alongside the colors (“Campfire”)
I am not sure if that is possible and feasible. For the date you probably just have to capitlize a “y” in the date format.
Dates can indeed be formatted.
For labels, as far as I know there is no way to control them separately. There are three options: keep them all, all the time; hide them all; and make them appear when the node is created, and fade later. This last option will make also appear the labels of the parent nodes, including the categories.
Unfortunately, at this scale and speed (1 day => 0.05 seconds) the third option still looks super messy. If you are visualizing one category (~ 2,000 posts, 1 day = 0.1 seconds) it is doable, though some of the paths looks really ugly (catname/subcatname/1234/5/6/7/8/9/10/11 or so).
Any feedback on the Edgeryders stories we could tell with these data? (“Hey, I can see when @johncoate started his monster “Status report” topic!” and similar)
Very cool! Looks like an 80s arcade game, or cell division.
I’m interested in parts where sections go from more shapeless forms to more self-contained, solidified communities. Does that always/only happen through the introduction of projects (e.g. Open Care, Spot the Future), or are there examples of it occurring more organically?
As for those projects themselves, are there moments where they solidify or increase more dramatically, and is it possible to understand why or how?
If nothing else, it’s definitely a cool visual that elicits an affective response from me. All these people coming together and creating stuff.
1 Like
I got a bit emotional. Thanks @alberto, it’s really wonderful.
Interesting choice of sountrack :-))
Especially because the active conversation are glittering so powerfully - with a different coloring of the background, it makes me think of the star system. We sometimes said we are working in ‘constellations’ - I think this fits nicely.
The things I noticed:
- the 2 big projects in 2014: when unMonastery drifted off and Spot the Future picked up
- the diversification in 2015: Nepal, Romania, Future Makers and beautiful stories…
- 2016 - Workspaces showing how much more specialized conversations became - you pointed it too in your description. But they float together somehow.
- 2018/2019 were also the years when they grow apart. I believe none of us could keep track anymore of everything going on. I see the constellations flowing freely at that time, it’s liberating in a way
2 Likes
More cell division, since 1980s arcade games looked like this:
Nodes are posts. They are added onto topics, which in turn are part of categories. In Edgeryders we are very action-oriented, so lively discussions tend to lead to “why don’t we do X?”. When a discussion becomes large, it receives its own category, and that’s when you see these “blossomings”.
That’s the holy grail, is it not? I guess we would need to start with smaller viz of individual projects (example), and make hypotheses: “We had a successful event in Warsaw in September, we should visually detect a “blossoming” of the conversation tree between August and October… look, there it is!”. Dataviz: telling stories with the data.
For you personally. @amelia, and the mighty @leonie and @ccs: we should be thinking about ways to use Gource to represent coding. This is not intuitive at all, for reasons connected to the high specificity of Gource. But for example, we could represent annotating as an edit to a node (supported in Gource: it shows a different-colored “laser beam”). To keep a visual record of the annotated node, we could create a second node representing the annotation, and assign to it a different color (this, would show also an extra green laser beam. Oh well.). In the end, the conversation graph would be a set of (say) green beads representing posts, interspersed by (say) blue ones representing annotations.
Another possibility is to represent annotations and codes as separate root branches. Instantiate codes as if they were topics; and annotations as if they were posts. Visually, you should see if some codes occur a lot (many annotations-nodes clustered around the same code-node). The main constraint is this: Gource can only deal with acyclical graphs (trees). So, we cannot watch the code-code network (or the social network) form in real time.
Yeah, it’s beautiful. 
1 Like
YouTube Studio off-the-shelf.
Thank you so much, this is exactly the input I was looking for!
1 Like
Frogger!
I was thinking more of Space Invaders and Asteroids 
But definitely more cell division.
1 Like
When we were working together at The Reef you were brute-forcing your graphics board to render in real time, which it could not do (n x 10ˆ4 objects to animate). In this case, I chose instead to output each frame to x264 codec. When you open the resulting file with an MP4 player, it runs smoothly.
1 Like
Cohesion is right now happening in my life, after @RoRemote nudged me to look at ER again, using the same SNA hooks as @martin (but I don’t think you two know each other).
Two questions for @alberto:
- Why did you chose nodes to be posts instead of people?
- Could posts be further categories by keywords (topic analysis, semantic analysis) to get a more fine grained view of knowledge and idea? In addition, what would be nice is to show how and when different ideas are merging into bigger ideas. Maybe there are ideas on how to approach this topic using the work on scientific creativity (e.g., Mukherjee & Uzzi (2016). A New Method for Identifying Recombinations of Existing Knowledge Associated with High-Impact Innovation: Innovation Combined with Existing Technology)
Relational event modeling (based on participation shift framework) could be used to potentially answer the question.
The work on team assembly might provide some ideas on what else could be done with that data.
Interesting work, I’d like if and how this will be continued.
The heavy lifting is done by Gource. What I did was to use Edgeryders data to generate a file that (1) represents online debate in a fairly intuitive way and (2) is digestible by Gource. The logic of Gource is that programmers are seen creating and editing files in a repository (with a tree structure), so what I did was mirror the logic. The files are replaced by posts; posts in a forum are also arranged in a tree (category=>subcategory=>topic=>post); and contributors are contributors.
As I progressed I was able to create visualization where annotations are visualized as well:
Indeed. Not sure how to represent this with Gource… yet.
2 Likes
I am still living in hope that I may one day experience Edgeryders visualised as the swarm of cats it is.
3 Likes