PyData Eindhoven 2019
2019-11-16..2019-11-19
Eindhoven, Netherlands
[PyData Eindhoven 2019, PyData Eindhoven 2020, PyData Eindhoven 2021, PyData Eindhoven 2022, PyData Eindhoven 2023, PyData Eindhoven 2024, PyData Eindhoven 2025]
who it's for
data scientists, machine learning engineers, MLOps engineers, data engineers, healthcare data scientists, research scientists, Python developers, data visualization specialists
topics
Thompson sampling and contextual bandits; propensity scoring and doubly robust estimation; variational autoencoders with KL loss; transfer learning from ImageNet; MLflow model registry and tracking; Apache Airflow DAG orchestration; Kubernetes operator for Airflow; process mining with pm4py; privacy and differential privacy techniques; secure multi-party computation (MPyC); image augmentation and Colab pipelines; uplift modeling for churn and domain adaptation
| title | author | length |
|---|---|---|
| Keynote: Personalization, Bandits, and Causal inference | Maurits Kaptein | 41m |
Sequential experimentation with bandit methods beats naive A/B testing: Thompson sampling and UCB rapidly reduce regret versus epsilon‑first. Contextual models… | ||
| Deep generative models for image... | Dimitra Gkorou, Koen Vannisselroij, Shama Khalil, Sonali Fotedar | 01h 17m |
Walks through building a convolutional Variational Autoencoder + KL loss in Keras/TensorFlow on the Fashion dataset (60,000 train / 10,000 test), trains (demo… | ||
| Managing Machine Learning Lifecycle with MLflow | Vladimir Osin, Milan Mulji | 57m |
Walks through MLflow setup, shows tracking server (SQLite backend + artifact store), records experiments from PyTorch training on Kuzushiji‑MNIST and logs… | ||
| Scheduling machine learning pipelines using Apache Airflow | Axel Goblet | 14m |
Shows how to schedule ML batch inference pipelines with Apache Airflow using per-attendee Airflow on AWS and Jupyter notebooks for DAGs. Demonstrates… | ||
| Image Recognition DL WORKSHOP | Maurice Peters | 39m |
An end-to-end workshop showing how to classify individual buildings from aerial and satellite imagery using a Google Colab notebook. It covers dataset layout… | ||
| Automated river plastic monitoring Using.. | Colin van Lieshout, Joery de Vos | 30m |
Claims that a camera plus deep learning pipeline can count and classify river plastics with practical accuracy and field deployments. Uses convolutional object… | ||
| Process Mining in Python | Sebastiaan J. van Zelst | 32m |
Process mining extracts process models and checks conformance from event logs using Python tools: a CSV/XES import pipeline into pandas, the pm4py library, an… | ||
| Preventing churn like a bandit | Gerben Oostra | 27m |
Companies lose 2–16% of monthly revenue to churn and the practical solution ties uplift to money: predict uplift directly (use the transform outcome trick)… | ||
| Transferring clinical prediction models across hospitals... | Mattia Fornasa | 23m |
Built an ICU readmission predictor called Padma critical on 14,000 Amsterdam admissions with AUC 0.8 then sees performance fall to 0.76 on later local data and… | ||
| Advertising, algorithms and Privacy: recent opensource developments | Ruben Mak | 28m |
Discusses concrete privacy wins and failures in advertising and algorithms: demonstrates reproducible de-anonymization of the Netflix Prize dataset, membership… | ||
| BigMedilytics: MPyC in practice | Natasja van de L'Isle | 22m |
Claims that secure multi‑party computation can train and evaluate clinical models without sharing raw records, demonstrated on a heart failure pilot with… | ||
| Airflow and Kubernetes at JW Player, a match made in heaven? | Rik Heijdens | 24m |
JW Player runs ~100 TB of analytics per month and replaces Airflow task code with Docker containers that run on Kubernetes via a custom operator that calls… | ||
| Common pitfalls leading to wrongly estimated model performance | Jan van der Vegt | 28m |
Common mistakes that make performance look better than it really is: using the validation set for model selection so validation becomes training and introduces… | ||
| The 3 energy saving tips for your data science ecosystem | Fahimeh Alizadeh | 20m |
Claims three concrete levers: monitor resources with AWS Cost Explorer and PSUtil to spot underused instances and schedule autoscaling or cron‑based… | ||
| Exploratory Data Analysis (EDA) and Visualization Techniques.. | Ekhtiar Syed | 24m |
Checks whether masks actually cover useful pixels not just counts uses pixel spacing to convert mask pixels to physical area and finds 12k chest x‑rays with… | ||
| Next-Gen railway asset management with... | Jasper Derikx, Richard Bartels | 28m |
Computer vision pipeline at ProRail locates insulation joints from a fleet of inspection images (about 50 million images per campaign), trains a small CNN… | ||
| Untitled12.ipynb | Vincent D Warmerdam | 35m |
Live demo shows messy notebooks turning into a reproducible pipeline by turning cleaning, feature engineering and aggregation into small functions and chaining… | ||
| A content-based recommender system for financial news | Anca Dumitrache, Feng Lu | 29m |
Content-based recommender for financial news that solves cold-start by modeling clicks as a per-user classification problem using implicit feedback and… | ||
| Pipelines for Fairness: A Convexing Usecase | Matthijs Brouns | 24m |
Pipeline approach presents concrete fixes and measured tradeoffs: a modified Gram-Schmidt transformer removes linear dependence on protected columns inside a… | ||
| A Python application to flag outliers in very high... | Satej Khedekar | 25m |
A Python app flags outliers in thousand‑dimensional wafer sensor vectors by projecting to low dimensions with UMAP then clustering with HDBSCAN, automatically… | ||
| Categorizing financial transactions for personal finance.. | Alexander Backus | 32m |
Describes how Yolt categorizes personal financial transactions to support personal finance management, focusing on the data and tooling used (Python ecosystem… | ||
| nothing matches | ||