-
-
Save kexline4710/7562581 to your computer and use it in GitHub Desktop.
| - Crime Prediction App (Nathan) | |
| - Crowdsourced Birthday Drink App (Ty) | |
| - Clinical Trial Patient Web App (Katy) | |
| - Annotate & Collect Links, maybe by categories (Meara) | |
| - Real Talk Conversations (Chirag) | |
| - Gunshot triangulation - mimicking data (Nathan) | |
| - If you have a bad day, see Nathan's face app (Ty) | |
| - Recreate AIM with the sound effects and look (Katy) | |
| - Secure chat (Nathan) | |
| - Visualize the tools/gems/languages in your github project (Katy) | |
| - Sponsor-a-person in need/buy a meal (Connor) | |
| - Respirator challenge.gov (Clay) | |
| - Crowd-source/equipment time-share 3D-printing, something else (Chirag/Connor) | |
| - Reddit for legislators (Connor) | |
| - Follow your leaders' twitter feed (Clay) | |
| - Open-source legislation / Third voice (Connor) | |
| - Tindr for babies (Clark) | |
| - Local/physical skillshare (Connor) |
+1 "If you have a bad day, see Nathan's face app" 😄
If you want to work with crime data, you should be aware of some recent history of what I'll call "crimey-neighborhood identification" apps. These usually breakdown like this:
Developer: Here's our civic app! It draws a red line around neighborhoods with high crime!
Press: Hooray! Technology! Woo!
Social justice activists: Are you @#&%*ing kidding me? It just drew lines around all the low income neighborhoods and told people never to go there. This is LITERALLY redlining.
Developer: DATA DON'T LIE! Technology! Woo!
A staggering diversity of racist shitbags now enter the conversation. 20 pages later...
Social data expert: Yes, data absolutely do lie. More specifically, crime data are created and modified discretionarily by police, which catalog crimes differently based on the race and class of perpetrators and particularly in intersection with the race and class of victims (ie black on white crime routinely draws harsher penalties than white on white or black on black crimes). There are several other vectors which create very high volume logging of events in neighborhoods with, for instance, police task forces assigned to them, that don't appear in other neighborhoods. This is well established in a number of studies. Data lie all the time, in that the digital representation in a database field also carries a lot of analog properties: how it was collected, by whom, and with what purpose. Putting disparate data sources into an aggregate like your app removes these analog properties of data, causing the appearance of certainty when none exists. The result is a technical artifact that is based on very weak evidence and happens to reinforce a century of destructive racial stereotypes. You should reconsider this project.
By now no one has seen this comment except the racist shitbags who respond with the comment "boobz".
If you want some insight on how to do this sort of thing better, here's a book I wrote on the subject. In short: if data don't tell you why something happened, they're not helpful data.
A few of us had an intense conversation about that this issue last night. Good point.
I have complete faith that you'll make amazing stuff.
Imagine how many ideas we would have if you cough included everyone.
Jonathan that was really insightful. thanks for sharing
Here are some of the links we got from open gov hack night too:
Open Gov Resources
Chicagoland Political Boundaries
The Redeye Homicide Map
And since we have a lot of open gov ideas on the list, here is a link to the Sunlight Foundation who already as a pretty big list of tools:
Sunlight Foundation Tools