PyData London - 69th meetup

Venue: 1 Angel Lane, EC4R 3AB

Please note:

1. A valid photo ID is required by building security. You MUST use your full real names on your meetup profile, otherwise, you will NOT make it on the guest list!
2. This event follows the NumFOCUS Code of Conduct, please familiarise yourself with it before the event.

Tickets are assigned through a lottery draw about 1 week before the event.

If your RSVP status says "You're going" you will be able to get in. No further confirmation required. You will NOT need to show your RSVP confirmation when signing in.

If you can no longer make it, please unRSVP as soon as you know so we can assign your place to someone on the waiting list.

***

Code of Conduct:
This event follows the NumFOCUS Code of Conduct, please familiarise yourself with it before the event. Please get in touch with the organisers with any questions or concerns regarding the Code of Conduct.

***

As always, there'll be free food & drinks, generously provided by our host, Man Group.

***

Main Talks
1️⃣ What's the Itch That Feature Stores Scratch? — Moritz Meister
Do you actually need a feature store for machine learning (ML)? Many individuals and organizations ask themselves this question, and many come to different conclusions. In this talk, we will clearly define how a feature store solves problems such as training-inference skew for operational ML systems, reuse of pre-computed features across multiple models, and the integration of historical and contextual data into online models. However, there are also scenarios where you do not need a feature store. For example, if you have a batch system processing data in a data warehouse and the features are reused in other models. We will try to give an honest appraisal of when you should reach for the feature store tool, and when to leave well alone.

2️⃣ Using BERT for Stable, Dynamic Topic Detection — Matt Hobby
BERT and other transformer models have revolutionised NLP and high dimensional representation of semantic meaning.
In this talk we discuss the use of BERT embeddings to detect topics in documents. We demonstrate that the combination of BERT embeddings with network community detection methods provides a stable means of detecting topics where the number of topics apriori is unknown, this overcoming one of major shortcomings of LDA.
We highlight the stability of the methodology when compared to the approach taken by the popular BERTopic GitHub repository. Furthermore, we demonstrate how this approach can be used for dynamic topic modelling.

Lightning Talks ⚡
🧀🧠 The Chemically Induced Christmas Experience — Emlyn Clay
Christmas can be a wonderful time of year, and often we have fond memories of the day spent with family and friends. However, as time wears on we can lose the Christmas spirit and perhaps we should look to Science to provide the answers.
I will take you on a whistle stop tour of the chemical and physiological processes involved in the perfect Christmas experience. You too can reclaim the innocence of youth. There will be numbers, there will be facts and you will be advised, strenuously, not to follow any content or advise within this talk.

Logistics
Doors open at 6.30pm (get there early as you have to sign-in via building security), talks start at 7pm, drinks from 9pm in the bar. We will have reduced capacity for this event but there will be plenty of people to discuss data science questions with!

Please unRSVP in good time if you realise you can't make it. We're limited by building security on the number of attendees, so please free up your place for your fellow community members!

Follow @pydatalondon (https://twitter.com/pydatalondon) for updates and early announcements.