Visual Culture Seminar
Design with Courage
Those were the prime words that stuck to me after having visited Postlight, a design studio in Manhattan NYC.
Data Visualization for Social Good
Aaron Hill — @aaronxhill — Parsons — The New School
what is datavis: hard to define
“JS, interaction, numbers” “view data in a new way in order to explore” “translating computer data in a human readable format” “infographic” “how to make data available” “document knowledge visually” “explain data without words” “analyzed data” “a way of telling stories” “making data easier to understand for people”
The moment you start working with data, you’re designing
DV & CD is doing the same thing with a different form of information
things like slideshows are dataviz, parsons course catalog is dataviz
Data doesn’t always fit into columns and rows
1850s london cholera outbreak map = dataviz
simple display of data can be extremely powerful
dataviz can be a way to convince people of certain data
process of understsanding the data that comes in is super important
the lost cat book map isn’t useful initially. In order to make it useful:
- Too much data / noise / invalid data
- Dealing with outliers
intend: what do you want to do and why acquire: get the data parse: convert the data to a useful format filter: get rid of outliers, aggregate dense data mine: statistics / machine learinng to extract meaning represent: represent the data visually refine: refine the visual representation interact: make it interactive publish: make it available to the world
computers are dumb, human parsing is necessary in order to make sure that the dataset is actually useful.
context clues can only be seen and used by humans
parse / filter is a lot of work
represent: software doesn’t matter, it’s about ways to display the data:
How should the data map to the visual representation?
Keep in mind that not every part of visual interface is in favor of transfering the data knowledge.
Jacques Bertin semiology of graphics:
basically any visualization can be reduced to points, lines, or areas.
Vox where we donate vs diseases that kill us chart is the worst chart ever made:
- Circles aren’t a good representation of data
- Data is not representable
Examples of good visualizations:
links: http://aaronhill.nyc/sxswdataviz.pdf http://lostcatbook.com/ http://aaronhill.nyc/historicdistrict/ https://rikghosh.github.io/thesis/ https://ny.curbed.com/2017/10/11/16461458/new-york-subway-map-massimo-vignelli-midtown
Timeframes are large.
Interactive data installation to balance the big corporation work.
Panorama photography is fun when it glitches (like when you’re on a moving train).
Project: create a huge panorama of the whole street of broadway
Doing panoramas has a lot of problems: stitching messes up perspective, time of day, different scenes over different days.
Different data layers are possible on the same broadway street: foursquare checkins, tweets, instagram posts, etc.
Cell reception on the MTA.
Being observed is as important as observing
Familiarity between actors increases willingness to share
With innocent objectives in gathering data, you oftentime get access to data you weren’t initially looking for.
The example given was the project where the movements of all the students were tracked. This resulted in the added ability to see what students slept with each other. Not data that they were looking for initially, but nonetheless data they got.
To what extent are we willing to give out personal information under the assumption it’s anonymous.
More from the type and interaction lady
Laying out found content in a new way.
Using personal websites for something else than personal work.
Exersize in style
A successful email will:
- Be black and white and use typography only
- Use your chosen found text
- Show form that is related to your content
- Be memorable, maybe surprising
This was a particularily useless exercise. I don’t understand why it’s going to help me in any way. I wish I didn’t come to this session.