What do different machine vision technologies do in fiction and art?
For the Machine Vision in Everyday Life project we’ve analysed how machine vision technologies are portrayed and used in 500 works of fiction and art, including 77 digital games, 190 digital artworks and 233 movies, novels and other narratives. You can browse the database we entered all our analyses in, or download the dataset as a set of .csv files from Dataverse – and you can read about the dataset in a data paper we just published over at Data in Brief.
This is my first experience publishing a dataset and a data paper and I have learned so much from the process! Learning R at the same time was brilliant – I’ve learnt how to do my own data analysis much more efficiently than I could using Excel and Gephi as I was before, and it’s just opened up so many new possibilities for what we could do with this data. I’ll certainly use this new knowledge when planning future projects. Writing a data paper was also a very useful process. Data in Brief have a very clear template, which was useful, and it was honestly quite interesting being forced to write a paper that needed to be purely descriptive, no analysis allowed. There’ll be other papers with analysis, don’t worry. When we wrote the paper we weren’t aware of other journals specifically for data papers, so another aspect of the process was figuring out how to fit humanities data into a journal that mostly publishes a different kind of dataset. Since then I’ve found the Journal of Open Humanities Data, which is specifically for the humanities, and I’m looking forwards to reading more there.
Right now I’m trying to finish up my book, Machine Vision: How Algorithms are Changing the Way We See the World, which will be published by Polity Press in late 2022 or early 2023. The manuscript is due August 1, so I’m writing and revising as hard as I can. So I shouldn’t really be thinking about my data right now…
But the thing about R is it’s really fun, especially when you have a dataset you’re really interested in, which of course I do since we just spent more than two years developing it. The fun of data analysis is not something I’d expected. I think it’s the instant gratification. You can have a question, like “do the different kinds of technologies do different things in machine vision situations, and you tinker around with some code, probably fail a few times and do some googling, and copy a snippet from here and from there, and then hey presto, you have a lovely visualisation like this!
I love how fast I learn (although I forget stuff constantly and have to look it up again), and I love how incremental it is. Perhaps that’s specific to data analysis. I want to make this graph or compute that data, and then once I’ve figured out how to that, I see what I still haven’t found out and realise that there’s something else I want to do so now I have to figure out how to do that.
What about drones, for instance? Are they portrayed as doing different things in different kinds of work?
Looks as though there are more drones in narratives (mostly novels and movies) than in games or digital art, anyway. And they’re mostly kept busy recording and killing. There aren’t that many drones in our dataset, though.
One of the challenges will be figuring out how to write about this data in an interesting way. I want to combine close readings and qualitative analysis with the data analysis, but that requires a new way of thinking about writing to what I’m used to.
Now I need to get back to writing my book…
Nick Montfort
Fascinating! The one piece of data I can’t understand: When you have a manuscript deadline of August 1, how is there a chance that your academic book will come out “in late 2022”!?
Jill
Ha! I suppose that is a bit optimistic – I actually haven’t asked when the book will be out. 2023 is probably more realistic!