Skip to content
Neural-net-powered web dashboard for a cafeteria
Python JavaScript HTML CSS
Branch: master
Clone or download

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.


Type Name Latest commit message Commit time
Failed to load latest commit information.

Cafeteria Dashboard

A fun little project to create a one-stop dashboard for a cafeteria, including crowd levels (current and historical averages - derived from webcam images by using a neural network), current weather, and today's menu.

Example Screenshot

How Does This Work?

Every minute from 10:30am to 2:00pm (when the cafeteria is open), a python script connects to the cafeteria webcam and downloads an image. The image is preprocessed (resized and converted to grayscale) and fed to an artificial neural network that was trained on some manually-labeled data. The neural network returns its best guess at how crowded the image looks (on a scale of 1-10), and this result is saved to a database. The dashboard queries the database and generates some lovely graphs.

Current weather conditions for the cafeteria location (to help the viewer to decide whether or not to sit outside) are pulled from the Weather Underground API.

Today's menu for the cafeteria is screen-scraped from the intranet site.


Install Anaconda:

Install Theano:

Install Curses:

pip install: requests requests_ntlm pillow theanets flask flask-sqlalchemy sqlalchemy-migrate simplejson beautifulsoup4

You can’t perform that action at this time.