PyData Exeter #10

Enjoy three talks from our brilliant speakers after drinks and networking.

Sponsors:
NumFOCUS
Exeter Innovation Hub
Butterfly Data - https://butterflydata.co.uk
Mekion - https://www.mekion.co.uk

Agenda:

• 18:45 Doors open
• 19:15 Talks start
• 20:30 Talks finish - stick around for a drink and networking at the CUCKOO
taproom across the road.

Talk #1 by Lucie Wright
From South West Water

Title: Natural Capital Mapping in the Water Sector

Right now we are experiencing a step-change in the way we manage the natural world around us, thanks to a combination of public datasets combined with high resolution satellite imagery now at our fingertips. This talk is an introduction to the real world application of an AI powered GIS system in the context of habitat mapping and how humans are essential to shaping its functionality.

Talk #2 by Jamie Conn and Matt Drew
From Sparx Learning

Title: Beyond multiple choice: Using LLMs for Marking and Feedback in Secondary Science
Sparx Science is an online homework system used weekly by over 400,000+ students in England, part of the broader Sparx portfolio which serves 2,600+ schools and 2.2m+ students.
Traditionally, open-response questions—requiring students to answer with anything from a few words to full paragraphs—have required marking by teachers. As a result, online homework systems typically relied on constrained question types, such as multiple-choice.
However, Large language models (LLMs) have opened the door to reliably mark and provide meaningful feedback on open-response tasks, supporting effective retrieval practice and challenging students to think deeply.
In this talk, we'll share our approach to using LLMs to automatically mark prose in Sparx Science, including: fine-tuning, prompt engineering, and everything in between.

Talk #3 by Yash Sakhuja
From Performalytics

Title: Presentation on how AI can be used for fuzzy matching
Showcasing how vector embeddings and unsupervised learning techniques can effectively deduplicate and merge messy customer data, illustrated through a real-world use case.

A little more about our speakers

• Jamie Conn

Leads the product development of Sparx Science. Formerly a Data Scientist developing personalisation algorithms used across Sparx products.

• Matt Drew

Data Scientist in the Modelling & Research team at Sparx. Works largely on the monitoring, development, and deployment of the various models used in our products.

• Yash Sakhuja

“I’m a Data & Insights Analyst at Performalytics, where I help e-commerce, retail, and sports brands unlock growth through data storytelling, machine learning, and AI. With a Master’s in Operations Research and Risk Analysis from the University of Manchester, I’ve delivered talks at PyData Manchester, Southampton, and Cambridge. Outside of work, I serve as the Vice President of Education at Toastmasters in Oxford, combining my love for public speaking with a passion for improv and all things data.”

CODE OF CONDUCT
The PyData Code of Conduct governs this meetup ([https://pydata.org/code-of-conduct/](https://pydata.org/code-of-conduct/)). To discuss any issues or concerns relating to the code of conduct or behaviour of anyone at the PyData meetup, please contact the PyData Exeter organisers, or you can submit a report of any potential Code of Conduct violation directly to NumFOCUS (https://numfocus.typeform.com/to/ynjGdT)