PyData meetup

26 January
  • Croy Castle, Aarle-Rixtel, the Netherlands
Live network event
PyData meetup

The global PyData network promotes discussion of best practices, new approaches, and emerging technologies for data management, processing, analytics, and visualization. This Meetup is a place for technical people to come and hear technical talks, and network with likeminded people in the Eindhoven region interested in data science.

We are excited to host a PyData meetup on January 26 at our EyeOn headquarters at Croy castle in the Netherlands.

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.

This Meetup is a place for technical people to come and hear technical talks, and network with likeminded people in the Eindhoven region interested in Python. No Sales, No Recruting, just technical talks.

 

This meetup will feature speakers from Pipple, Ikea, and EyeOn: 

 

Pipple: Is there a next step in warehouse logistics optimization?

Speakers:

  • Wouter Leenen: Operational Director
  • Lennart van Ham: Data Scientist

Combining operations research heuristics with deep reinforcement learning

The growth of online web shops over the past decade has been accompanied by the arrival of many e-fulfillment centers. An incredibly dynamic operation, if only because of the short delivery times. Ordered today, delivered tomorrow.

Increasing mechanization and automation are necessary given the tight labor market and consequently rising personnel costs. The result: an extremely dynamic and complex process. Human intelligence can no longer manage these operations efficiently and optimally.

Heuristics have been used in Operations Research (OR) for decades and are capable of making quick and – in general – good decisions. But are they adaptive enough as well?

With its worth already proven in complex gaming environments, Reinforcement Learning (RL) learns through a simulated version of reality. As such, it is able to anticipate constant changes in a dynamic environment, without having to re-calculate a thing. But a warehouse is no game of chess, and is RL cut out to fully grasp its immense complexity?

We believe we can get the best of two worlds. Real-time response in a dynamic environment whenever required while reducing complexity using heuristics where they are best suited. Not replacing, but combining OR-heuristics with innovative Machine Learning technologies.

In this talk, we will introduce the dynamics in warehouse logistics, show how RL can be used to accelerate OR-techniques and give practical examples of applications in real-world scheduling problems.

Learn more about Pipple!

 

Ikea: Organizing order fulfillment using mathematical optimization

Speakers:

  • Victoria Guerrero Mestre: Data Scientist
  • arturo pérez rivera: Data and Analytics leader

The increasing popularity of e-commerce brings new selling opportunities for retailers. However, selling online also brings new logistical challenges. In this talk, we will discuss a particular challenge for IKEA: organizing the fulfillment of e-commerce orders. We will give an overview on how fulfillment decisions are made today using business rules and how they will be made tomorrow using mathematical optimization models and data analytics. We will illustrate how the new decision paradigm is implemented (as a microservice in Python) and what potential value it can bring to our customers and operations.

 

EyeOn: Driver-based demand forecasting: beyond existing frameworks

Speaker:

  • Mark Ramakers: Data Scientist

Time series can be tricky to handle, especially when there are hundreds of thousands of them, like in supply chain demand forecasting. Forecasts are essential for an efficient supply chain in order to make informed decisions and optimize inventory and customer service levels. Despite the potential of machine learning in this area, it is not a simple task. In fact, it has only been in the past few years that machine learning has been able to outperform traditional statistical methods in this context (source). During this talk, we will explore the application of machine learning for supply chain demand forecasting, as well as our own unique framework that can be used to achieve this goal.

Search for