show and tell for hackers in DC.

Round 36: Unfair Witty Outburst

13 Sep 2016

jessica garson

Burning Man Missed Connections

http://twitter.com/jessicagarson

http://www.burningmanmissedconnections.com

https://github.com/JessicaGarson/JessicaGarson.github.io

Have you ever been to Burning Man? It’s kind of like a weeklong festival for rich hippies to take drugs and be promiscuous.

Jess compiled the best of the Craiglist Missed Connections from Burning Man.

Uses the static site generator Pelican, with markdown for each post. Pelican is a good way to get started with blogging about programming. It’s written in Python and has a simple CLI.

Pelican helpful for creating simple websites (E.g. About Me page, Technical blog)

Jess scrapes Craigslist and then manually skims through for the best stories

Jess teased a tantalizing story about being invited to travel to Burning Man on a private jet. Catch up with her about this at the bar…

Ben Robinson

DC Open Education Data Shiny Application

https://twitter.com/benj_robinson

Ben patiently endured the group singing happy birthday.

learndc.org

Resources for parents, community members, and educators

School profiles section has data on DC schools – graduation rate, attendance rate, enrollment, race.

All of this data is available; the site has an API

Clean the data up and make it more friendly to the average user

LearnDC is an R package and web tool to use LearnDC’s public API about schools: https://github.com/benjaminrobinson/LearnDC

R has a web development framework called Shiny; so he created a Shiny app with his brother

https://benjaminrobinson.shinyapps.io/learndc_shiny_downloader/

Data available:

Graduation data, standardized testing data, attendance, teacher information, enrollment, suspensions, special education, and more

You can examine the data in a JS-based web app and also download the data (flat file).

Jared Nielsen

Cityscape of DC

http://twitter.com/jarednielsen

I built a cityscape of DC using 3D printed & laser cut components using FOSS. He only had 30 days to do this and he had his little brother help him. He outsourced the 3D printing work to a team in Fairfax, did the laser-cutting himself.

Used Blender for 3D modeling to make a totally amazing cityscape. The cityscape is meant to depict the D.C. area in “the future.” His models aren’t based on realistic proportions/measurements because he didn’t want to have to give rights to anyone.

Local motors is a company that 3d-prints cars. They have a storefront in National Harbor. The goal was for Local Motors to teach kids how to 3D print models and inspire them to work on 3D printing.

They have a soft lab where you can 3D print things; they wanted to be able to draw people in so Jared created a 3D-printed and laser cut city scape.

The model includes both DC and the VA sides of the “diamond” (Arlington and Fairfax Co.).

All of this was done in 30 days – and since they were so short on time, Jared had to

They recorded the whole installation on a GoPro camera and it made for an incredible time-lapse video! ABConf

Eric Schles

Document search with elasticsearch and ocropy http://twitter.com/EricSchles

Shannon Turner

Free yourself from the burden of your secrets

http://twitter.com/svthmc

http://shannonvturner.com/secret

Shannon’s app allows people to input secrets (in text form) and receive an image with shapes and colors determined algorithmically based on the text. It produces a unique image for every secret.

The algorithm produces a one-way hash. It outputs a hexadecimal string. All of the processing is done on the client side; it doesn’t go to the server (safe from the NSA!).

If someone has the same secret as you, you’ll receive a notification. (does this include some sort of fuzzy matching for punctuation/spelling, etc.?).

Jim Webb

Which Republican Are You?

https://whichrepublican.com/

Good facial recognition software is proprietary but there’s a relatively new software OpenFace that works well

Most facial recognition systems measures the distance between 15 or 16 different points; this one uses a neural network that was trained on 500k people to understand which points are the most important.

Uses a training set of about 20-40 images per Republican to match.

Torch does an analysis of the faces uploaded and calculates the distance between the points.

Running on Webfaction

This baby looks like Chris Christie.

It takes a few seconds to process the facial recognition, so what’s next is creating a queue for the pictures and telling people about how DC doesn’t have voting representation while they wait!