PyDataMCR September

PyDataMCR September Talks

Continuing to kick this year off with some more great talks we are hosted this month by Krakenflex.

THE TALKS

Unlocking climate data for SME carbon accounting - Joss Bennett
**"**99% of UK businesses are SMEs, and around three-quarters are sole traders… they account for 43-53% of UK businesses’ greenhouse gas emissions in aggregate” (Source: UK Finance). Net zero presents a real challenge for SMEs in the UK, with tough an economic landscape and growing regulatory requirements which is taking the form of cascading requirements for climate disclosure originating from large multinationals down the supply-chain.

Increasing tools to automate data analysis, which overcome traditional hurdles of admin time and quality to reporting have emerged. Focusing on a spend-based assessment to identify areas for data improvement and ultimately collecting supplier-specific data will encourage action over paralysis to enable effective value-chain decarbonisation.

Joss Bennett, Sustainability, Energy and Carbon Reporting Lead for [Perse.io](http://perse.io/). With 6 years of career operating across the UK sustainability landscape Joss has driven environmental campaigns with organisations such as Blake Made and Treeapp, respectively. This cumulated in a founding associates position, as the Lead Carbon Consultant, for Manchester based (UK) sustainability software start-up [NetZero.co.uk](http://netzero.co.uk/).

Joss was a Senior Consultant at Green Economy, part of the social enterprise The Growth Company, implementing Greater Manchester Combined Authority’s decarbonisation targets. He has recently accepted a position at Perse.io to enable the widespread automation of utility data for systemic decarbonisation across the UK and beyond.

Rapidly Prototyping RAG applications using Local Models - Saurav Maheshkar
Saurav will discuss the process of making performant RAG applications using LlamaIndex and local models. Walking through an example use-case: scraping data from the internet, creating a local embedding and LLM model, generating a vector index, and running queries. We will also discuss LLMOps best practices and create a quick UI for the application using Gradio and Hugging Face Spaces.

Saurav Maheshkar is a Machine Learning Engineer at LightlyAI (Ex: Re:course AI, Weights & Biases). He’s also a Google Developer Expert in Machine Learning and is interested in Geometric Deep Learning and Representation Learning. He’s passionate about open source and enjoys delivering talks.

Small Data - Dealing with Climate uncertainty - Josh Hayes

Climate change is impacting everyone, and business activities are one of the primary sources of emissions, financial institutions need to take responsibility for their share of these emissions.

The Taskforce for Climate-related Financial Disclosures (TCFD) have set regulatory requirements for financial institutions to disclose their financed emissions.
The requirements for disclosures is tricky, as lenders need to understand their customers' emissions. As many businesses do not disclose their own emissions, a range of methodologies must be used.

Josh will explore the data challenges faced by financial institutions, provide an explanation of the standard methodologies used to estimate finances emissions, and introduce the idea of data quality scores, a metric that can be used to indicate the quality of an emissions estimate in lieu of a quantified confidence interval.

Josh has a PhD in astrophysics from the University of Manchester in 2021, where he was using Bayesian modelling to explore exoplanetary atmospheres. After completing his thesis, he worked at NatWest, where he has been the technical lead behind the calculation of the bank's emissions. He is leading research streams into significant methodological improvements, whilst supporting on multiple additional climate initiatives.

LOCATION
We'll be at Krakenflex, who are also kindly supplying catering. The capacity is limited to 90.

After the talks we'll all head somewhere local for some post-event socialising.

EVENT GUIDELINES
PyDataMCR is a strictly professional event, as such professional behaviour is expected.

PyDataMCR is a chapter of PyData, an educational program of NumFOCUS and thus abides by the NumFOCUS Code of Conduct

https://pydata.org/code-of-conduct.html

Please take a moment to familiarise yourself with its contents.

ACCESSIBILITY
Under 16s welcome with a responsible guardian. There is a quiet room available if needed. Toilets are accessible!

SPONSORS
Thank you to NUMFocus for sponsoring Meetup and further support
Thank you to AutoTrader for sponsoring PyDataMCR.
Thank you to Krakenflex for sponsoring PyDataMCR, as well as hosting this event!