This web page is meant to inform about the research project Cafeteria Coordination, which involves data collection in the canteen of the Max Planck House Tübingen.
This project will hopefully (1) help reduce waiting times and (2) allow us to study the problem on a theoretical level, including time series forecasting and game theory, within the research project Cafeteria Coordination.
The web app "Cafload" with current queue length and preliminary daily forecast is available at cafload.tue.mpg.de. For privacy reasons, the web app is only accessible from within the campus. Here's a screenshot of what it looks like.
In what follows, you find news as well as a general description of the project and system.
19.3.2018: Part of our research in the Cafeteria Coordination project is now summarized in this arXiv preprint.
23.2.2018: Beginning of this week, a simple beta version of a daily updated forecast went online at cafload.tue.mpg.de.
29.9.2017: Web page cafload.tue.mpg.de with live canteen queue length information goes officially online. (So far, due to privacy constraints, this web page is only accessible for members of the campus.) (updated 8.12.207)
4.7.2017: We are now working on publishing (an estimate of) the current queue length on this web page, so that everyone on the campus can access this information. We will inform the campus once this feature is available.
19.4.2017: We added a seventh sensor recently. Transfer and storage of (anonymized) data are up and running now. Furthermore, our algorithm for inference of the current queue length seems to work (approximately), although we have to run more and longer tests.
9.2.2017: The sensors are in operation for some time now. We are currently working on configuration of the sensors, as well as transfer, storage and preprocessign of the data, which will take a while.
18.1.2017: The sensors were installed yesterday. The data collection will start as soon as some details have been figured out.
Short description of the project
Cafeteria load web application
Current queue information is available at cafload.tue.mpg.de. Keep in mind:
Currently, the cafeteria load information is only accessible from campus IPs. Soon we will also provide access via a password.
Anonymized web access data (anonymized IP and limited browser information) is collected on this web page for scientific purposes.
The research project involves data collection in the canteen of the Max Planck House.
Here are the most important facts:
The primary goal is to model waiting times in the cafeteria, and to develop instruments that help to reduce these waiting times.
To make this possible, some of the customers' anonymized data needs to be collected.
To ensure everyone's privacy, no personal data is collected.
There are seven stereoscopic sensors installed in the waiting and self-service area of the cafeteria to count heads and to measure the waiting time. No video is saved or transmitted outside the sensor system.
The electronic payment system transmits when and where (at which counter) payments happened. No further information is revealed, especially no names and card IDs.
All work councils of the Max Planck Campus Tübingen as well as privacy officers of the Max Planck Society were involved in the setup of the project.
Thank you for your understanding!
Persons responsible: Philipp Geiger, Bernhard Schölkopf.
Data privacy statement (Datenschutzerklaerung): available for download (note that soon there will probably be a minor update).