diff --git a/pydata-eindhoven-2021/category.json b/pydata-eindhoven-2021/category.json new file mode 100644 index 000000000..a420cf8e6 --- /dev/null +++ b/pydata-eindhoven-2021/category.json @@ -0,0 +1,3 @@ +{ + "title": "PyData Eindhoven 2021" +} diff --git a/pydata-eindhoven-2021/videos/an-incomplete-list-of-implementing-data-science-effectively-erick-webbe-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/an-incomplete-list-of-implementing-data-science-effectively-erick-webbe-pydata-eindhoven-2021.json new file mode 100644 index 000000000..e7e303cf1 --- /dev/null +++ b/pydata-eindhoven-2021/videos/an-incomplete-list-of-implementing-data-science-effectively-erick-webbe-pydata-eindhoven-2021.json @@ -0,0 +1,40 @@ +{ + "description": "The talk is aimed at everyone that wants to learn how to put data science in practice more effectively. The traits can help individual contributors, collaborating teams or even organisations that are looking to become more effective in how they organize. Sharing these traits will hopefully speed up your development as a data science enthusiast.\nThe talk will touch on the following topics:\n-Brief intro of Data Science @ bol.com\n-Introduce the six traits that made the list (so far)\n-Highlight two in more details from personal experience and actual use cases\n-Engage with the audience for feedback and reflection\n\nIn the highlighted examples I will talk about two use case from personal experience. In the first, I'll talk about how we introduced a paradigm shift in how we create and use data driven forecasts to improve how we operate many processes across bol.com. In the second, I'll share the approach taken to help governmental agencies make better use of data to predict COVID outbreaks and enable them to adopt their strategy using novel techniques and data sources.\nThe following traits will be presented and highlighted (in rough order):\n-Fail fast to learn fast\n-Understand your solution\n-Pick the right tool\n-Take small directed steps\n-Collaborate with T-shaped teams\n-Excite your users\n\nThe talk will be supported with real time feedback from Menti and encourages feedback from the audience. This list was compiled based on many iterations and lots of feedback, and will definitely change after this session once again. I'm hoping you'll help and further refine the list for future sessions.\n\nErick Webbe\nLinkedIn: https://www.linkedin.com/in/erick-webbe-2634a633//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1845, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.linkedin.com/in/erick-webbe-2634a633//", + "url": "https://www.linkedin.com/in/erick-webbe-2634a633//" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Erick Webbe" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/CbgLFnuMidc/maxresdefault.jpg", + "title": "An incomplete list of implementing data science effectively", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=CbgLFnuMidc" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/applied-ai-on-the-edge-maurits-kaptein-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/applied-ai-on-the-edge-maurits-kaptein-pydata-eindhoven-2021.json new file mode 100644 index 000000000..7ff408b35 --- /dev/null +++ b/pydata-eindhoven-2021/videos/applied-ai-on-the-edge-maurits-kaptein-pydata-eindhoven-2021.json @@ -0,0 +1,36 @@ +{ + "description": "While the potential of Machine Learning (ML) and Artificial Intelligence (AI) is widely recognized in various sectors (health, industry, commerce, etc.), regretfully many ML/AI projects do not make it past the Proof of Concept (PoC) stage. In this talk I will share a number of my own experiences with \u201cfailed\u201d AI projects (i.e., projects that easily passed the PoC stage, but never made it into production), and I will examine the root causes of these failed AI projects. To do so, I will have to provide a bit more background regarding the various types of AI models/projects that exists, explain how AI works in some detail, and discuss the common production/deployment patterns that companies use in their attempts to scale their ML/AI activities. Effectively, I will describe the AI deployment process from data collection, to AI model development, to model evaluation, and finally towards large scale model deployment. At each of these stages I will highlight the challenges involved and the common points of failure.\nNext, I will turn my talk to potential solutions: although it is hard to find a uniform solution for scaling every possible ML/AI solution in the book, for a large class of applied AI/ML models efficient and effective deployment methods have recently been developed. I will explain how deploying AI models on the edge (i.e., not in the cloud) solves a number of common AI deployment problems. Furthermore, I will explain how modern technological advances enable the effective deployment of trained ML/AI models on edge devices despite the diversity in device types (e.g., different hardware, different computational constraints). Finally, I will argue that deployment on the edge makes applied AI more scalable, reduces the energy footprint of AI, improves user privacy, and reduces operational costs of many AI applications.\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2137, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Maurits Kaptein" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/Cp3d_FqoLYA/maxresdefault.jpg", + "title": "Applied AI on the edge", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=Cp3d_FqoLYA" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/be-kind-to-yourself-spend-less-time-on-data-exploration-willem-hendriks-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/be-kind-to-yourself-spend-less-time-on-data-exploration-willem-hendriks-pydata-eindhoven-2021.json new file mode 100644 index 000000000..36bc962bc --- /dev/null +++ b/pydata-eindhoven-2021/videos/be-kind-to-yourself-spend-less-time-on-data-exploration-willem-hendriks-pydata-eindhoven-2021.json @@ -0,0 +1,48 @@ +{ + "description": "Is data exploration the most important step in a data project? Weird how this crucial phase almost never gets the credit its deserves. \nI will do an attempt here, a little ode to the exploration phase, and try make it cool again by using some modern packages. Any tool that helps us being effective, increases the change we find that gem during data explorations. I believe any insights gained during the data exploration phase pays back at least double later on in the journey.\n\nWillem Hendriks : Studied mathematics, and working with data since graduation. From small statistical sets to using big data tooling, currently at Big Data Republic.\nGitHub: https://gitlab.com/whendrik/\nLinkedIn: https://www.linkedin.com/in/willemhendriks//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:08 Introduction\n2:05 What is Data Exploration?\n2:26 \"You have to get to know your data!\"\n11:50 Pandas Profiling \n14:07 SweetVIZ\n15:59 DABL \n17:58 dTreeViz\n19:56 dtale \n22:35 Conclusions\n25:18 Q&A\n\nS/o to https://github.com/arne-cl for the video timestamps! \n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1798, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://gitlab.com/whendrik/", + "url": "https://gitlab.com/whendrik/" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://www.linkedin.com/in/willemhendriks//", + "url": "https://www.linkedin.com/in/willemhendriks//" + }, + { + "label": "https://github.com/arne-cl", + "url": "https://github.com/arne-cl" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Willem Hendriks" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/tiNQDY8ixXU/maxresdefault.jpg", + "title": "Be kind to yourself! Spend (less) time on Data Exploration", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=tiNQDY8ixXU" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/biosensor-machine-learning-with-julia-matthijs-cox-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/biosensor-machine-learning-with-julia-matthijs-cox-pydata-eindhoven-2021.json new file mode 100644 index 000000000..714fb8d76 --- /dev/null +++ b/pydata-eindhoven-2021/videos/biosensor-machine-learning-with-julia-matthijs-cox-pydata-eindhoven-2021.json @@ -0,0 +1,52 @@ +{ + "description": "Biosensors allow you to perform data science on your own bio signals, like brainwaves, heart rates and electrical muscle signals. How cool is that?! In this talk I'll briefly introduce you to the open source community growing around biosensor devices. \nI will explain briefly what biosensors are. What kinds are available to consumers and engineers. What sensors do I have myself? How to obtain the data from them. \nFor the latter part I will introduce the open source BrainFlow library, for which I am a voluntary open source developer. It is a fast performing c++ library that's easy to use, can be deployed anywhere and has many language bindings, including Python and Julia. BrainFlow allows anyone with a biosensor to extract their own data with little hassle and build applications in their favorite programming language.\nI will show code examples for how to obtain data from such biosensors, using Julia and BrainFlow. I specifically demonstrate gesture predictions using myo-electric muscle sensors. This topic I investigated for building open source bionic arm control systems. I would like to do a live demo, or at least show videos of the live data streaming and machine learning predictions. You can control real devices with these algorithms, like bionic arms, but I have also used it to play video games.\nThe live data processing, data streaming and machine learning predictions need to be as fast as possible, so I wrote all my code in Julia to show it is performant enough for this task.\n\nMatthijs Cox : A physicist who loves to code. I taught myself data science, software development and neurotechnology. I learned all about biosensors and neurotechnology as founder of the Symbionic Project where we tried to create more affordable bionic arms.\nCurrently I am a product architect at ASML, developing numerical computing applications. We struggle a lot with the two language problem at ASML. In order to solve this problem I discovered Julia about two years ago. Julia is now my favorite programming language.\n\nGitHub: https://github.com/matthijscox/\nTwitter: https://twitter.com/matthijscox/\nLinkedIn: https://www.linkedin.com/in/matthijscox//\nWebsite: https://www.functionalnoise.com/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1569, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.linkedin.com/in/matthijscox//", + "url": "https://www.linkedin.com/in/matthijscox//" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://github.com/matthijscox/", + "url": "https://github.com/matthijscox/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.functionalnoise.com/", + "url": "https://www.functionalnoise.com/" + }, + { + "label": "https://twitter.com/matthijscox/", + "url": "https://twitter.com/matthijscox/" + } + ], + "speakers": [ + "Matthijs Cox" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/kLj0JQKhNMM/maxresdefault.jpg", + "title": "Biosensor Machine Learning with Julia", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=kLj0JQKhNMM" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/combining-imitation-reinforcement-learning-to-win-the-bot-bowl-competition-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/combining-imitation-reinforcement-learning-to-win-the-bot-bowl-competition-pydata-eindhoven-2021.json new file mode 100644 index 000000000..b4bdcce49 --- /dev/null +++ b/pydata-eindhoven-2021/videos/combining-imitation-reinforcement-learning-to-win-the-bot-bowl-competition-pydata-eindhoven-2021.json @@ -0,0 +1,48 @@ +{ + "description": "This paper describes a hybrid agent trained to play in Fantasy Football AI which participated in the Bot Bowl III competition. The agent, MimicBot, is implemented using a specifically designed deep policy network and trained using a combination of imitation and reinforcement learning. Previous attempts in using a reinforcement learning approach in such context failed for a number of reasons, e.g. due to the intrinsic randomness in the environment and the large and uneven number of actions available, with a curriculum learning approach failing to consistently beat a randomly paying agent. Currently no machine learning approach can beat a scripted bot which makes use of the domain knowledge on the game. Our solution, thanks to an imitation learning and a hybrid decision-making process, consistently beat such scripted agents. Moreover we shed lights on how to more efficiently train in a reinforcement learning setting while drastically increasing sample efficiency. MimicBot is the winner of the Bot Bowl III competition, and it is currently the state-of-the-art solution.\n\nNicola Pezzotti: is a Senior Scientist in Artificial Intelligence at Philips Research, Eindhoven, the Netherlands and assistant professor at Eindhoven University of Technology. His research interests include machine learning, medical imaging, visual analytics, explainable AI, optimization techniques and software engineering. He received his BSc and MSc degrees in Computer Science and Engineering from the University of Brescia, Italy, in 2009 and 2011. He received his PhD cum Laude from Delft University of Technology, the Netherlands, in 2018. Besides his experience in the startup world, he was a visiting scientist to INRIA Saclay, Paris, in 2017 and Google AI, Zurich, in 2018. He is recipient of several awards, including the IEEE VGTC Best Dissertation Award, TU Delft Excellence in Research and the Dirk Bartz Prize for Visual Computing in Medicine.\n\nGitHub: https://github.com/Nicola17/\nTwitter: https://twitter.com/nicolapezzotti/\nLinkedIn: https://www.linkedin.com/in/nicola-pezzotti-58670527/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1749, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.linkedin.com/in/nicola-pezzotti-58670527/", + "url": "https://www.linkedin.com/in/nicola-pezzotti-58670527/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://twitter.com/nicolapezzotti/", + "url": "https://twitter.com/nicolapezzotti/" + }, + { + "label": "https://github.com/Nicola17/", + "url": "https://github.com/Nicola17/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Nicola Pezzotti" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/XX-OkBcG5WI/maxresdefault.jpg", + "title": "Combining Imitation + Reinforcement Learning to win the Bot Bowl competition", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=XX-OkBcG5WI" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/compressive-sensing-iman-mossavat-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/compressive-sensing-iman-mossavat-pydata-eindhoven-2021.json new file mode 100644 index 000000000..db6c9b8bb --- /dev/null +++ b/pydata-eindhoven-2021/videos/compressive-sensing-iman-mossavat-pydata-eindhoven-2021.json @@ -0,0 +1,40 @@ +{ + "description": "One can regard the possibility of digital compression as a failure of sensor design. If it is possible to compress measured data, one might argue that too many measurements were taken. - David Brady Compressive sensing main idea is to measure and compress to cope with the scarcity of resources. For example the limited resource can be battery power or limited communication band-width in simple sensors or measurement time in magnetic-resonance imaging. This talk first explains what it means to measure in a compressed mode and then how we can use prior knowledge about the structure of the signals to measure them in a compressed mode. Finally it shows how the prior knowledge used in compressive sensing moved forward over the few years from simple sparsity constraints to more advanced assumptions about the manifold of measurements. Such manifolds can be constructed in a data-driven manner for example by deep-learning. The talk briefly touches upon the statistical physics of inference and uses compressive sensing as a case study where information and computational complexity meet.\n\nIman Mossavat\nI am a machine learning expert with more than 10 years of experience in applications related to challenging inverse problems. Currently, I am looking into lowering the energy needed to power AI.\n\nMy interests beyond machine learning are emergent phenomena in complex networked systems such as social media networks, and the brain. I believe that information processing and energy consumption are closely connected via notions such as free-energy, and that the study of complex self-organizing natural phenomena can teach us tremendously in this regard.\n\nIn the past years, my focus was on the development of signal processing algorithms for (optical) metrology applications (Tools that measure at nanometer accuracy). (Bayesian) probabilistic modeling, machine learning and numerical optimization are tools I regularly use.\n\nLinkedIn: https://www.linkedin.com/in/mossavat/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1764, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.linkedin.com/in/mossavat/", + "url": "https://www.linkedin.com/in/mossavat/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Iman Mossavat" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/pFqZkdgWHo0/maxresdefault.jpg", + "title": "Compressive Sensing", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=pFqZkdgWHo0" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/dealing-with-the-versioning-of-production-ready-models-corne-vriends-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/dealing-with-the-versioning-of-production-ready-models-corne-vriends-pydata-eindhoven-2021.json new file mode 100644 index 000000000..0d6fb7269 --- /dev/null +++ b/pydata-eindhoven-2021/videos/dealing-with-the-versioning-of-production-ready-models-corne-vriends-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "In this talk we will start with what we do at Eneco and how we operate at scale with our smart thermostat Toon\u00ae. Furthermore, we will discuss how Databricks and Python allow us to serve predictions from our models at scale. The good and the bad of this approach and how maintaining this setup has its own set of obstacles.\nEspecially, we will delve into our specific problem that we encounter with maintaining our models that are in production. Instead of focusing on the more model related aspects (e.g. model drift) we focus on the operational aspect to make sure that the environment is not as old as when the first iteration of the model was made. This ensures that the latest innovations from Databricks (and Spark) flow through to our environment that Data Scientists and Data Engineers rely upon.\nWe will cover several possible solutions to this problem, from considering a portable format, such as ONNX, to complete containerization of each of the models in production using MLFlow Projects. In the end, we will showcase, what in our specific case is to our knowledge the most pragmatic solution. This talk is addressed to anyone who is interested in the more operational aspects of ML at scale.\n\nCorn\u00e9 Vriends: I am a Data Scientist that works at Eneco, specifically on the Toon\u00ae.\n\nLinkedIn: https://www.linkedin.com/in/cornevriends//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n0:00 - Welcome!\n0:09 - Introduction\n1:12 - Problem Statement\n2:38 - Company Introduction\n5:23 - Design Architecture\n7:16 - Machine Learning Flow\n8:56 - Pickling\n13:29 - Solving Pickling Issue\n14:26 - Possible Solutions\n15:50 - 1. Portable Format (ONNX)\n17:56 - 2. Containerize\n19:50 - 3. Periodically Retrain\n22:28 - Key Takeaways\n23:35 - Questions\n\nS/o to https://github.com/yugant10-commits for the video timestamps!\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1951, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://www.linkedin.com/in/cornevriends//", + "url": "https://www.linkedin.com/in/cornevriends//" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://github.com/yugant10-commits", + "url": "https://github.com/yugant10-commits" + } + ], + "speakers": [ + "Corné Vriends" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/vSKgGHUE7Rw/maxresdefault.jpg", + "title": "Dealing with the versioning of production-ready models", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=vSKgGHUE7Rw" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/estimating-solar-production-w-distributed-deep-learning-from-smart-meter-data-eindhoven-2021.json b/pydata-eindhoven-2021/videos/estimating-solar-production-w-distributed-deep-learning-from-smart-meter-data-eindhoven-2021.json new file mode 100644 index 000000000..c8adedb96 --- /dev/null +++ b/pydata-eindhoven-2021/videos/estimating-solar-production-w-distributed-deep-learning-from-smart-meter-data-eindhoven-2021.json @@ -0,0 +1,54 @@ +{ + "description": "Smart meters only measure redelivered energy, but can\u2019t measure how much solar energy was used by the household itselff before that. In this talk we will go in-depth into the process of estimating solar production from the energy that is redelivered to the grid at the smart meter, using PyTorch and Horovod. We will go through the entire process from ideation to production-ready: validating the feasibility of the model, gathering a dataset, exploring different solutions and finally training and validating the model. \nNext to that there are some technical challenges: To deal with the scale of data we used the distributed deep learning framework Horovod for training, and Spark for distributed predictions. We show how we use MlFlow for keeping track of our model iterations and finally, we will walk you through the data-engineering required to get the data presented in our mobile app! \nCome join us to learn the ins and outs of getting a deep learning use case to production at scale!\n\nRik van der Vlist: A Data Scientist with a passion for making things work from beginning to end, I love the challenges that come with maintaining, deploying and automating machine learning pipelines just as much as the modelling part itself. At Eneco we have developed a lot of interesting models for smart energy advice, ranging from estimating the savings of lowering your setpoint by one degree to detecting inefficient washing machines or monitoring solar panels.\n\nCorn\u00e9 Vriends: I am a Data Scientist that works at Eneco, specifically on the Toon\u00ae.\nLinkedIn: https://www.linkedin.com/in/cornevriends/\n\nFenno Vermeij: I am a Data engineer that works for Eneco. My interests include AWS, Azure, Spark and big, big data. I am working hard to learn more about architecture. I am working on Toon, trying to make smart homes better.\nLinkedIn: https://www.linkedin.com/in/fenno-vermeij-3193a831/?originalSubdomain=nl\nGitHub: https://github.com/fennovj/\nWebsite: https://fenno.dev/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2190, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://fenno.dev/", + "url": "https://fenno.dev/" + }, + { + "label": "https://www.linkedin.com/in/cornevriends/", + "url": "https://www.linkedin.com/in/cornevriends/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.linkedin.com/in/fenno-vermeij-3193a831/?originalSubdomain=nl", + "url": "https://www.linkedin.com/in/fenno-vermeij-3193a831/?originalSubdomain=nl" + }, + { + "label": "https://github.com/fennovj/", + "url": "https://github.com/fennovj/" + } + ], + "speakers": [ + "Rik van der Vlist", + "Corné Vriends", + "Fenno Vermeij" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/3LbkXCGl1Gk/maxresdefault.jpg", + "title": "Estimating solar production w/ distributed deep learning from smart meter data", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=3LbkXCGl1Gk" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/explainable-causal-inference-results-thomas-nagele-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/explainable-causal-inference-results-thomas-nagele-pydata-eindhoven-2021.json new file mode 100644 index 000000000..72a26f75c --- /dev/null +++ b/pydata-eindhoven-2021/videos/explainable-causal-inference-results-thomas-nagele-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "A Bayesian network is a graph representation of a joint probability distribution over a number of variables of interest. Once instantiated with prior probabilities and observations, the inference algorithm updates all probabilities for all variables, i.e., nodes in the graph. In large networks it is difficult to understand what observations are relevant for the inferred probability of a node of interest. To come to the set of observations relevant to a node of interest, we compute an extended Markov Blanket for this node. This approach provides to the user a subgraph with only the relevant observations, and, therefore, supporting the explanation over the inferred probabilities.\nThe talk consists of 1) an introduction into Bayesian networks and some of the available BN Python libraries, 2) a description of our method to make the inference results more explainable, and 3) some examples, both showing the method\u2019s capabilities and limitations. The talk does not require any prior knowledge, but a rough understanding of probability theory and graphs could help. After the talk, the audience will be familiar with Bayesian networks and how these can be scoped based on relevance.\n\nThomas N\u00e4gele: a research fellow at ESI (TNO) in Eindhoven. His focus is on the development of model-based methodologies for system-level diagnostics of high-tech systems. His main interests are model-based methodologies, model simulation and multi-model methods such as co-simulation. He received his BSc, MSc and PhD degrees in Computing Science from the Radboud University in Nijmegen.\nLinkedIn: https://www.linkedin.com/in/thomas-n%C3%A4gele/\nWebsite: https://thomasnagele.nl/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1981, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://www.linkedin.com/in/thomas-n%C3%A4gele/", + "url": "https://www.linkedin.com/in/thomas-n%C3%A4gele/" + }, + { + "label": "https://thomasnagele.nl/", + "url": "https://thomasnagele.nl/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Thomas Nägele" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/iqEri-q_sUU/maxresdefault.jpg", + "title": "Explainable causal inference results", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=iqEri-q_sUU" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/how-to-quickly-build-data-pipelines-for-data-scientists-geert-jongen-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/how-to-quickly-build-data-pipelines-for-data-scientists-geert-jongen-pydata-eindhoven-2021.json new file mode 100644 index 000000000..7d25826a1 --- /dev/null +++ b/pydata-eindhoven-2021/videos/how-to-quickly-build-data-pipelines-for-data-scientists-geert-jongen-pydata-eindhoven-2021.json @@ -0,0 +1,48 @@ +{ + "description": "Data pipelines usually consist of loading the data, transforming it and writing to some other location. Initially, this does not sound very complicated. The question arises why it is so hard then to do? In this talk we will discuss how to perform these steps in pyspark, and especially what the latest developments are around delta lake, data quality checks and data modeling. What patterns are preferable and why? At the end of this talk data engineers and data scientists should have a view on a pattern that will fit in a lot of general situations and will help them to set up a pipeline more quickly while preventing a lot of issues upfront.\n\nGeert: is a data consultant working at Pipple, with extensive experience in the domain of data engineering, data science and software. Developing data platforms using cloud native technologies is what he enjoys most. Especially the messy process of bringing POCs into production is what he likes to do.\nGitHub: https://github.com/Jongen87/\nTwitter: https://twitter.com/Jongen87/\n\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Introduction\n01:58 Data Pipelines\n07:32 Tip 1: Define a Clear Split Between Data Engineering and Data Science\n08:58 Tip 2: For the Data Pipeline Part Use Notebooks in Flows\n10:34 Tip 3: Have a Sit-Down With Your Team and Decide on Standards\n11:14 Tip 4: All Data You Use Needs to Have a Source\n12:36 Tip 5: Prepare for Changes\n13:36 Delta Lake\n16:38 Demo\n26:50 Conclusion\n\nS/o to https://github.com/mraxilus for the video timestamps!\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1846, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://github.com/Jongen87/", + "url": "https://github.com/Jongen87/" + }, + { + "label": "https://twitter.com/Jongen87/", + "url": "https://twitter.com/Jongen87/" + }, + { + "label": "https://github.com/mraxilus", + "url": "https://github.com/mraxilus" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Geert Jongen" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/XMnDCZhm9Go/maxresdefault.jpg", + "title": "How to quickly build Data Pipelines for Data Scientists", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=XMnDCZhm9Go" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/mlops-101-tips-tricks-and-best-practices-vladimir-osin-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/mlops-101-tips-tricks-and-best-practices-vladimir-osin-pydata-eindhoven-2021.json new file mode 100644 index 000000000..b7b9b6759 --- /dev/null +++ b/pydata-eindhoven-2021/videos/mlops-101-tips-tricks-and-best-practices-vladimir-osin-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "Machine Learning Operations is the most hyped topic these days, but let's have a talk without much hype. Shall we? In this one, I am going to discuss stages of the machine learning project with emphasis on practical aspects of each stage. In these 30 minutes you will learn about scoping, data, modelling and deployment parts of the ML project. Proper understanding of these stages will allow a smooth transition of your Untitled_final.ipynb to production deployment bringing value to your business stakeholders. The focus of this talk is to guide the audience throughout the ML lifecycle considering best practices and advice for the following stages:\n-Scoping \n-Data\n-Modeling \n-Deployment \n\nBy the end of this talk, you will know which important questions to ask before starting ML project, answers on which will boost the development cycle. We will also briefly touch upon tooling for MLOPS and ML team composition. The takeaway of tips and best practices with help you to set current and future ML projects on the rails of success.\n\nVladimir Osin\nGitHub: https://github.com/osin-vladimir\nLinkedIn: https://www.linkedin.com/in/vosin//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2034, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://github.com/osin-vladimir", + "url": "https://github.com/osin-vladimir" + }, + { + "label": "https://www.linkedin.com/in/vosin//", + "url": "https://www.linkedin.com/in/vosin//" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Vladimir Osin" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/dzSp3Zf897g/maxresdefault.jpg", + "title": "MLOPS 101: Tips, Tricks and Best Practices", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=dzSp3Zf897g" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/optimal-on-paper-broken-in-reality-vincent-d-warmerdam-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/optimal-on-paper-broken-in-reality-vincent-d-warmerdam-pydata-eindhoven-2021.json new file mode 100644 index 000000000..57047cd95 --- /dev/null +++ b/pydata-eindhoven-2021/videos/optimal-on-paper-broken-in-reality-vincent-d-warmerdam-pydata-eindhoven-2021.json @@ -0,0 +1,52 @@ +{ + "description": "In particular, I'll discuss:\n-a general method to calm down optimistic claims of optimal results\n-how easy it is to draw the wrong lesson when running a grid-search\n-how data is a much better proxy for interpretation than hyperparameters\n-how easy it is to find wrong labels in public datasets\n-why sentiment models are generally a bit strange\n-tricks to deal with some of these issues\n\nThere will also be a demo, and an announcement, of a new python project\n\nMy name is Vincent, ask me anything. I have been evangelizing data and open source for the last 7 years. You might know my from tech talks where I attempt to defend common sense over hype in data science.\nCurrently, I work as a Research Advocate at Rasa where I collaborate with the research team to explain and understand conversational systems better.\n\nIn my spare time, I maintain a suite of open source packages (scikit-lego, human-learn, whatlies, tokenwiser, clumper, memo, pytest-duration-insights, mktestdocs, and more). I'm also building some online services (calmcode.io and dearme.email).\nGithub: https://github.com/koaning/\nTwitter: https://twitter.com/fishnets88/\nLinkedIn: https://www.linkedin.com/in/vincentwarmerdam//\nWebsite: https://koaning.io/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2194, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://twitter.com/fishnets88/", + "url": "https://twitter.com/fishnets88/" + }, + { + "label": "https://koaning.io/", + "url": "https://koaning.io/" + }, + { + "label": "https://github.com/koaning/", + "url": "https://github.com/koaning/" + }, + { + "label": "https://www.linkedin.com/in/vincentwarmerdam//", + "url": "https://www.linkedin.com/in/vincentwarmerdam//" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Vincent D. Warmerdam" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/lJKPiOf_o8k/maxresdefault.jpg", + "title": "Optimal on Paper, Broken in Reality", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=lJKPiOf_o8k" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/real-time-transaction-categorization-w-bayesian-feedback-loop-tijl-kindt-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/real-time-transaction-categorization-w-bayesian-feedback-loop-tijl-kindt-pydata-eindhoven-2021.json new file mode 100644 index 000000000..0fc3a4a7d --- /dev/null +++ b/pydata-eindhoven-2021/videos/real-time-transaction-categorization-w-bayesian-feedback-loop-tijl-kindt-pydata-eindhoven-2021.json @@ -0,0 +1,40 @@ +{ + "description": "As part of our personal finance management functionality, ING categorises transactions in real-time for customers that want to get insight into where they are spending their money. In this talk for anybody who would like to be inspired by a real-life, large-scale data science use case, I\u2019ll elaborate on 1) the tools and constraints within which we are operating, 2) how we bootstrapped the algorithm, so we could go live to customers with acceptable quality, and 3) the Bayesian-inspired functionality that we created to make this categorisation algorithm self-learn from the feedback from those customers.\nThe system, which was developed in-house, is split into a batch part and a real-time part. The batch part was written in PySpark and uses both implicit and explicit feedback from customers to tweak the predicted category for each counterparty. We use Bayesian inference to assess probabilities and bandits to provoke more varied feedback from customers. The real-time part uses the information from the batch part to categorise the transactions as they happen using Apache Flink and PMML.\nIn this talk I hope to inspire you to make the most of the tools you have available, and get your hands dirty with streaming analytics and to convince you that you don\u2019t need deep learning to change the lives of your customers for the better.\n\nTijl Kindt: a senior data scientist at ING Bank in the Netherlands. After obtaining a Bachelor's degree in Astronomy, and a Master's degree in Media Technology with a thesis on predicting mood in the context of movie recommendations, he joined ING in 2013, where he worked for 3 years on marketing models with SAS. Since then he's been focusing on customer-facing data products that help customers manage their money, such as Kijk Vooruit (Look Ahead) and Inzicht (Insight).\n\nLinkedIn: https://www.linkedin.com/in/tijlk/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2355, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://www.linkedin.com/in/tijlk/", + "url": "https://www.linkedin.com/in/tijlk/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Tijl Kindt" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/nmHtl2YAMVY/maxresdefault.jpg", + "title": "Real-time transaction categorization w/ Bayesian feedback loop", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=nmHtl2YAMVY" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/responsible-implementation-of-machine-learning-models-for-life-and-death-decision-making-eindhoven.json b/pydata-eindhoven-2021/videos/responsible-implementation-of-machine-learning-models-for-life-and-death-decision-making-eindhoven.json new file mode 100644 index 000000000..c95009f56 --- /dev/null +++ b/pydata-eindhoven-2021/videos/responsible-implementation-of-machine-learning-models-for-life-and-death-decision-making-eindhoven.json @@ -0,0 +1,56 @@ +{ + "description": "An increasing amount of AI-driven software products is being productionised in hospitals and used by doctors to make critical decisions. To ensure the responsible implementation of ML models, it is vital that models give adequate outputs not only in the curated training environment, but also in the messy world of real-time data.\nThis is especially the case in health care, where hundreds of different vital signs, laboratory values, medications and patient demographics may contribute to ML predictions. Changes in data registration processes, manual data entry errors and large changes in patient population or treatment policies happen regularly and unexpectedly. \nWhat can we do to ensure responsible implementation of ML given these challenges? Pacmed Critical is one of the first examples where AI is used in production in Dutch hospitals. The software supports intensive care doctors, by predicting whether patients can be safely discharged from the ICU. In this talk we present the three-fold way Pacmed Critical ensures responsible predictions in production:\nWhen (not) to predict? How the ML model avoids giving predictions when it should not be trusted, using both business rules and out-of-domain detection\nHow to ensure data validity? How ML monitoring enables automatic continuous monitoring of real-time data problems and data drift \nWhy did that prediction change? How a dashboard which visualises the change of Shapley values during a patients' admission, to interpret and explain the causes of prediction changes for an individual patient.\n\nMichele : is a Senior Data Scientist and Machine Learning Engineer at Pacmed, where he works on developing AI-powered decision support tools for personalized medicine, in particular for Intensive Care Units. His academic background is in Medical Engineering and Robotics, and he has experience working in deep learning research and fin-tech. At Pacmed, he focuses on building scalable and interpretable pipelines to train and implement explainable models in production.\nGithub: https://github.com/michetonu/\nTwitter: https://twitter.com/MicTonu/\nLinkedIn: https://www.linkedin.com/in/micheletonutti//\nWebsite: https://michetonu.github.io/about.html/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 - Welcome!\n00:10 - Introduction\n01:06 - Bridging the implementation gap in healthcare\n02:34 - Agenda\n02:55 - Pacmed Critical and the Intensive Care\n06:44 - Challenges of implementing ML for life-and-death decisions\n07:36 - Data Validity\n14:35 - Interpretability and Auditability\n19:10 - Prediction Validity\n24:24 - What happens when there is a problem in production?\n26:54 - Next steps\n28:42 - Conclusions\n30:06 - Q&A\n\nS/o to https://github.com/michetonu for the video timestamps!\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2074, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/michetonu/", + "url": "https://github.com/michetonu/" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://www.linkedin.com/in/micheletonutti//", + "url": "https://www.linkedin.com/in/micheletonutti//" + }, + { + "label": "https://twitter.com/MicTonu/", + "url": "https://twitter.com/MicTonu/" + }, + { + "label": "https://michetonu.github.io/about.html/", + "url": "https://michetonu.github.io/about.html/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://github.com/michetonu", + "url": "https://github.com/michetonu" + } + ], + "speakers": [ + "Michele Tonutti" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/r63oijinKhI/maxresdefault.jpg", + "title": "Responsible implementation of Machine Learning models for life-and-death decision making", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=r63oijinKhI" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/some-attention-for-attenuation-bias-ruben-mak-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/some-attention-for-attenuation-bias-ruben-mak-pydata-eindhoven-2021.json new file mode 100644 index 000000000..3a7c756a5 --- /dev/null +++ b/pydata-eindhoven-2021/videos/some-attention-for-attenuation-bias-ruben-mak-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "The outcomes of our models are not always what we intent to get. In my opinion, we should pay more attention to attenuation bias (also referred to as measurement error bias or regression dilution). When there is noise in the independent variables (i.e. features) the parameters of your model will be biased towards 0. You might think this is only an issue when doing inference, but from a machine learning perspective you might suffer the exact same problems depending on how the predictions are being used.\nI will explain orthogonal regression, which is traditionally used to solve attenuation bias. I will use this example to explain why you can only correct for attenuation bias when having at least some information about the noise in your independent variables.\nThe second part will be about how we handle attenuation bias in geo experiments at our company. I will first introduce geo experiments for causal inference and explain why there is potential attenuation bias. We will then dive into the code to show how we can account for this attenuation.To my knowledge, our method is not used outside of our company. Resources like Google's white paper do refer to orthogonal regression, but do not mention the simple but powerful solution we propose.\nAs a bonus, I would like to explain how this relates to imputation for missing variables. I would then like to argue why you should be careful in following Andrew Gelman in applying random regression imputation because of attenuation bias.\n\nRuben Mak: Co-founder of PyData Eindhoven and presented a couple of times at PyData Amsterdam. Fan of Bayesian statistics, causal inference and differential privacy. Working with my amazing colleagues on cutting-edge technologies, defining the future of the advertising ecosystem.\nGitHub: https://github.com/rubenmak \nLinkedIn: https://nl.linkedin.com/in/rubenmak/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1877, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://nl.linkedin.com/in/rubenmak/", + "url": "https://nl.linkedin.com/in/rubenmak/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://github.com/rubenmak", + "url": "https://github.com/rubenmak" + } + ], + "speakers": [ + "Ruben Mak" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/bq7_PB0L_k4/maxresdefault.jpg", + "title": "Some Attention for Attenuation Bias", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=bq7_PB0L_k4" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/synergizing-model-and-human-decision-making-using-augmented-machine-learning-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/synergizing-model-and-human-decision-making-using-augmented-machine-learning-pydata-eindhoven-2021.json new file mode 100644 index 000000000..394f71cd3 --- /dev/null +++ b/pydata-eindhoven-2021/videos/synergizing-model-and-human-decision-making-using-augmented-machine-learning-pydata-eindhoven-2021.json @@ -0,0 +1,48 @@ +{ + "description": "For our credit risk modelling, we love to use machine learning models to support human decision-making. The past two years we worked hard towards making our models more transparent to support our risk department. We also democratised our data by making it available to everyone in the company.\nIn my talk, I will introduce you to our approach of augmenting our machine learning and fuelling our feedback loop to continuously improve on our features and modelling approaches. Elements will include:\n-Co-designing a business rule model together with domain experts \n-Increasing transparency by sharing feature descriptions with our predictions\n-Gathering feedback by providing model explanations with Shapley values\n-How models and the risk department collaborate using challenges\n\nKay: is a Data Scientist / Machine Learning Engineer at Floryn, a FinTech based in 's-Hertogenbosch, the Netherlands. At Floryn he builds, deploys and monitors models to predict the financial health of companies. He likes to find pragmatic solutions to complicated problems and enjoys managing projects to bring value in a structured manner.\nGitHub: https://github.com/kayhoogland/\nLinkedIn: https://www.linkedin.com/in/kay-hoogland//\nWebsite: https://www.kayhoogland.nl/\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 2048, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://www.linkedin.com/in/kay-hoogland//", + "url": "https://www.linkedin.com/in/kay-hoogland//" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://github.com/kayhoogland/", + "url": "https://github.com/kayhoogland/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://www.kayhoogland.nl/", + "url": "https://www.kayhoogland.nl/" + } + ], + "speakers": [ + "Kay Hoogland" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/9p4HILGG_XY/maxresdefault.jpg", + "title": "Synergizing model and human decision-making using augmented machine learning", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=9p4HILGG_XY" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/the-journey-from-home-brew-analysis-scripts-to-a-product-andrew-rutgers-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/the-journey-from-home-brew-analysis-scripts-to-a-product-andrew-rutgers-pydata-eindhoven-2021.json new file mode 100644 index 000000000..d752897e4 --- /dev/null +++ b/pydata-eindhoven-2021/videos/the-journey-from-home-brew-analysis-scripts-to-a-product-andrew-rutgers-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "Most engineers dabble in programming, hate repeating things, and like optimising, so they usually have a few directories full of python or Matlab scripts to help them with their analysis. While these tools can be very powerful, they are rarely understandable, let alone usable to anyone else. But what if those scripts can be turned into a product? Andrew has built a few tools from own-analysis scripts into products for a wide range of users. He will discuss some of the challenges of building software for other people, and some of the approaches including using Lambda functions, API Gateway, React and a variety of other tools.\n\nAndrew Rutgers: Andrew is the CEO of ChargeSim a SaaS Simulation and analysis tool which helps plan EV chargers for fleets, cutting the costs to change bus and other fleets to electric. He started with electric vehicles racing solar powered cars, then moved on to solar-powered airplanes. He has built software throughout his career, including in C and Assembly for 8-bit microcontrollers for power electronics, C++ for image analysis during his PhD, Matlab for analysis tools, recently he has been working in Python and React and is Julia-curious. He has a PhD in Electrical Engineering and an MBA in Circular Economy. In his spare time he boulders, runs marathons and co-hosts the Tangible Computing Podcast\nGitHub:https://github.com/andrew-chargesim/\nLinkedIn: https://www.linkedin.com/in/andrew-rutgers//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1935, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://www.linkedin.com/in/andrew-rutgers//", + "url": "https://www.linkedin.com/in/andrew-rutgers//" + }, + { + "label": "https://github.com/andrew-chargesim/", + "url": "https://github.com/andrew-chargesim/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Andrew Rutgers" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/2sHKPYLIAhw/maxresdefault.jpg", + "title": "The journey from home-brew analysis scripts to a product", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=2sHKPYLIAhw" + } + ] +} diff --git a/pydata-eindhoven-2021/videos/when-performance-matters-running-a-real-time-computer-vision-pipeline-pydata-eindhoven-2021.json b/pydata-eindhoven-2021/videos/when-performance-matters-running-a-real-time-computer-vision-pipeline-pydata-eindhoven-2021.json new file mode 100644 index 000000000..1f5f460e1 --- /dev/null +++ b/pydata-eindhoven-2021/videos/when-performance-matters-running-a-real-time-computer-vision-pipeline-pydata-eindhoven-2021.json @@ -0,0 +1,44 @@ +{ + "description": "We have been developing an automated camera solution, in which we use video input, neural networks for player detection, sport-specific decision making and a PTZ-camera to deliver an automated livestreaming application. The complete pipeline has to work in real-time with a maximum delays of 300ms. Within the 300ms we have to fetch the video streams, run several neural networks, apply business rules and control the PTZ. Over the past year, we have had to refactor code, remove complete modules or change python packages due to delay constraints. During the talk we will show the insights we gained during the process. Specifically, we will go into:\n(a) overall architecture of the pipeline.\n(b) system requirements.\n(c) implementing a logging module to better identify the bottlenecks in the applications.\n(d) use high performance queues for communication.\n\n\nJelmer Wilhelm: Background in industrial engineering, data analysis, machine learning and artificial intelligence. Currently working at Whitebox Data Science; a company specialized in custom software and data analysis for logistics companies. Also working at Eyedle; a company specialized in computer vision software.\nGithub: https://github.com/j-wilhelm/ \nLinkedIn: https://www.linkedin.com/in/jelmer-wilhelm//\n\nPyData Eindhoven 2021\nWebsite: https://pydata.org/eindhoven2021/\nTwitter: https://twitter.com/pydataeindhoven\n\n===\n\nwww.pydata.org\n\nPyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. PyData provides a forum for the international community of users and developers of data analysis tools to share ideas and learn from each other. The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. PyData communities approach data science using many languages, including (but not limited to) Python, Julia, and R. \n\nPyData conferences aim to be accessible and community-driven, with novice to advanced level presentations. PyData tutorials and talks bring attendees the latest project features along with cutting-edge use cases.\n\n00:00 Welcome!\n00:10 Help us add time stamps or captions to this video! See the description for details.\n\nWant to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/numfocus/YouTubeVideoTimestamps", + "duration": 1803, + "language": "eng", + "recorded": "2021-11-12", + "related_urls": [ + { + "label": "Conference Website", + "url": "https://pydata.org/eindhoven2021/" + }, + { + "label": "https://twitter.com/pydataeindhoven", + "url": "https://twitter.com/pydataeindhoven" + }, + { + "label": "https://www.linkedin.com/in/jelmer-wilhelm//", + "url": "https://www.linkedin.com/in/jelmer-wilhelm//" + }, + { + "label": "https://github.com/numfocus/YouTubeVideoTimestamps", + "url": "https://github.com/numfocus/YouTubeVideoTimestamps" + }, + { + "label": "https://github.com/j-wilhelm/", + "url": "https://github.com/j-wilhelm/" + }, + { + "label": "https://pydata.org/eindhoven2021/", + "url": "https://pydata.org/eindhoven2021/" + } + ], + "speakers": [ + "Jelmer Wilhelm" + ], + "tags": [], + "thumbnail_url": "https://i.ytimg.com/vi/CCC7mrrFlpQ/maxresdefault.jpg", + "title": "When performance matters: running a real-time computer vision pipeline", + "videos": [ + { + "type": "youtube", + "url": "https://www.youtube.com/watch?v=CCC7mrrFlpQ" + } + ] +}