Hello everybody. Welcome to the October PyData meetup.
A big thank you to our host, Wooga, for hosting us and for providing food and drinks for everyone.
Marc Tonsen and Pablo Prietz
Pupil - An Open Source Eye Tracking Platform powered by Python
The analysis of human gaze behavior is a long-standing research field with applications in many areas, including psychology, medicine, marketing, and human-computer-interaction. Eye movements are commonly measured by means of mobile eye trackers. Such devices are head-worn and feature small cameras for recording images of the eyes, which are analyzed in order to deduce gaze information.
Pupil Labs is a Berlin-based company that was founded in 2014. Today it is one of the world-leading manufacturers of mobile eye trackers. Pupil Labs stands out due to their commitment to building an eye tracking pipeline that is entirely open-source.
In this talk, we will give you an introduction to eye tracking and the Pupil Labs software stack, which is build almost exclusively in Python.
Marc Tonsen obtained his masters degree in Computer Science from Saarland University. His thesis focused on novel eye tracking methods. He joined Pupil Labs in 2016 as the first member of the R&D team in 2016. He is currently a Partner and R&D Lead.
Pablo Prietz obtained his bachelors degree in Cognitive Science at the University of Osnabrueck. He joined Pupil Labs in 2016. He is currently the Software Engineering Lead at Pupil Labs and manages a diverse open-source community. In parallel to his work at Pupil Labs he is pursuing his masters degree in Computer Science at TU Berlin.
Exploring RESTful APIs in the Notebook
Restful APIs, although not strictly well-defined, are the bread and butter for exposing microservices or services in general (also for ML) on the internet today. It’s impossible to avoid them, but adopting a new API within the modern paradigm of exploratory computing can still be painful. This talk will present an emerging prototype solution that goes beyond curl and friends and aims at making it easier and faster to explore Restful APIs without leaving your beloved Jupyter notebook environment. The prototype builds on ideas known from Postman, and stands on the shoulders of giants like the requests and ipywidgets packages. Being nearly ready for others to use, one purpose of this example-driven presentation is, therefore, finding others who might want to help getting it done faster and better. Imagine all the plugins you could write!
Dinu Gherman has studied computer science at Karlsruhe University. He worked in domains like mobile telephony systems, financial reporting solutions, documentation and IoT platforms for companies in various countries and sizes from small to large. He is now working for HERE Technologies on the second generation of various services available on the still emerging Open Location Platform. Aside from that he takes a certain pride in having ignited and contributed to building the local Python community in Berlin.
Maciej Gryka (Lightning talk)
Snakes of the world, unite! Modern Python workers seizing Production.
In this talk, I will briefly describe how we used Docker, AWS Batch and EKS to create a modern and scalable setup for long-running Python workers for data tasks. I will touch briefly on testing and CI setup and then show how all the pieces fit together and what are advantages and the trade-offs involved. At the end I will show our own application to illustrate what kinds of problems it's a good fit for.
Maciej leads the science team at RainforestQA trying to create intelligent human-in-the-lopp systems to make Quality Control easy and fast. Before joining Rainforest as the first data scientist, he was working on eye tracking algorithms and completed a PhD in Computer Vision and Machine Learning before neural networks were cool.
If you want to give a lightning talk, send us an email at [masked] or come and say hi.