talk-data.com

Venue: Carnival House, 100 Harbour Parade, Southampton, SO15 1ST 📢 Want to speak 📢: submit your talk proposal

Main Talks 1️⃣ Searching for Shady Patterns: Shining a light on UK corporate ownership - Adam Hill In June 2016 the UK government launched the world’s first “beneficial ownership” register; a requirement for all UK companies to register who were the “persons of significant control”, PSCs, who actually controlled the company. Recent investigative journalism has made headlines with the leaking of the Panama and Paradise papers and it is clear that transparency in corporate ownership needs to be a significant factor within modern democracy. In a partnership between DataKind UK and Global Witness we have built the worlds first network graph mapping all of the UK public data on those who control corporate interests in the UK; it comprises in excess of 4.5 million companies and 4 million individual people. It has been enriched with company officer data and metrics of financial secrecy based upon geographic regions.

The goal of the project was to enable Global Witness to search for "shady patterns" within corporate ownership networks to act as leads for investigative journalism to expose corrupt practices. Further more, we were able to analyse the completeness of the register and identify ways of improving such data structures to inform other world governments how to best build similar public registers of corporate ownership.

We present here how we built this amazing data structure using Python tools for cleaning and data processing and a Neo4j graph database storing the network graph itself. In addition, we share the first insights derived from this process.

2️⃣ Breaking the Black Box - How to Evaluate Your Agents... in Real Time Too! - Craig West If you are building with LLMs, creating high quality evaluations is one of the most impactful things you can do. Without evals, it can be very difficult and time intensive to understand how different model versions might affect your use case. This talk aims to provide you a roadmap that may be simpler than you think to implement.

In this talk, we will look at the two aspects of Observability and Evaluation. Using the manual evaluating-ai-agents.com, along with its code repo, we will see that observability can be done without vendor solutions but with standard Python, either during Evaluation Driven Development or after development.

We will look at three core evaluation strategies - deterministic, human and LLM as Judge - with code examples.

Lightning Talks ⚡ ⚡1️⃣ TBA ⚡2️⃣ TBA

Please note:

1. 🚨🚨🚨A valid photo ID is required by building security. You MUST use your initial/first name and surname 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.

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. *** There will be pizza & drinks, generously provided by our host, Carnival UK. ***

Logistics Doors open at 6.30 pm, talks start at 7 pm. For those who wish to continue networking and chatting we will move to a nearby pub/bar for drinks from 9 pm.

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 @pydatasoton (https://twitter.com/pydatasoton) for updates and early announcements. We are also on Instagram/Threads as @pydatasoton, and find us on LinkedIn.

This is an IN-PERSON meeting

Speaker: Adam Machanic

Abstract:

For many database professionals, crafting an optimal query is half guesswork and half black art. If you switch the order of the WHERE predicates while burning sandalwood incense, will the query suddenly move a bit quicker? Or maybe you need to use a CTE—that will help, right? While speculative methods like these can eventually yield success, it doesn't come without excessive hand wringing and time wasting. But it doesn't have to be this way! In this session we'll start by revisiting the foundations. You’ll learn about the key logical guarantees on which query optimization is based, and why a rewrite that doesn't bend a guarantee is almost never going to help. From there we'll investigate various query patterns and take a hard look at some commonly suggested tuning advice. Are subqueries a problem? Do you need to worry about join order in your query? What about the oft-maligned IN predicate? All of this will be illustrated with various examples in both SQL Server and PostgreSQL. By the end of the session, you'll understand which rewrites might work, in which scenarios, and—most importantly—which you can totally ignore. And as a bonus, you might even be able to retire your incense collection.

Aside from victory questions like "can black force a checkmate on white in 5 moves?" many novel questions can be asked about a game of chess. Some questions are trivial (e.g. "How many pieces does white have?") while more computationally challenging questions can contribute interesting results in computational complexity theory. In this episode, Josh Brunner, Master's student in Theoretical Computer Science at MIT, joins us to discuss his recent paper Complexity of Retrograde and Helpmate Chess Problems: Even Cooperative Chess is Hard. Works Mentioned Complexity of Retrograde and Helpmate Chess Problems: Even Cooperative Chess is Hard by Josh Brunner, Erik D. Demaine, Dylan Hendrickson, and Juilian Wellman 1x1 Rush Hour With Fixed Blocks is PSPACE Complete by Josh Brunner, Lily Chung, Erik D. Demaine, Dylan Hendrickson, Adam Hesterberg, Adam Suhl, Avi Zeff