Leeds Data Science Meetup will be holding their September event during Leeds Digital Festival this year. Our regular events cover the cutting-edge work in data science and data engineering that is happening in and around Leeds.
You can meet with students, scientists, engineers, and entrepreneurs who are interested in all things Data Science related! This is an exciting time for coders of every type in Leeds, so join us!
In September, we have speakers from both academia and industry. Professor Roy Ruddle (LIDA) will be presenting “How should I investigate data quality?” and Myles Mitchell (Jumping Rivers) will be presenting “Demystifying MLOps”.
MLOps is an approach to data science where the use of trained models in production is a key focus.
If we consider how a model changes during the evolution of a project, there are many factors involved. Incorporating additional training data, changes to the model structure, or simply retraining an existing model, will cause model outputs to change. Keeping track of the different versions of a model, and making it easy to deploy new versions and roll back to earlier versions of a model is critical. Conversely, the performance of a trained model may start to wane over time, such that monitoring of a deployed model is necessary.
In this talk, Myles will discuss some open-source tools that he has been investigating that simplify using MLOps in practice.
Do you know that there are more than 100 ways in which data can be of “low” quality? That is one reason why data preparation often takes more than half of a data science project’s time.
In this talk, Roy will show you how to investigate data quality in an efficient yet rigorous manner. He will demonstrate the approach with an openly available six-step workflow and associated Python package.
Principal Data Scientist, Jumping Rivers
Professor of Computing (University of Leeds) and Research Technology Director (LIDA)
