Search – talk-data.com

Title & Speakers	Event
Analysing PMQs with Power BI, Python and data science (almost) - Ned Stratton 2025-02-25 · 18:00 Hansard online archives all debates in the UK Parliament going back over 100 years. The set-piece event of this is arguably PMQs, a cantankerous weekly 30-minute televised head-to-head between the Prime Minister and the Leader of the Opposition. Since Tony Blair took office in 1997, PMQs debate transcripts have generated nearly 5 million words. Turning this vast bank of unstructured text data into a focused and rich analytical dataset to produce insights into trends in UK politics sounds advanced and 'data-sciency', right? Do we need sentiment scoring and word2vec? An LLM? No, as it happens. With some subject-matter affinity and mostly vanilla Python (and some help from its text library NLTK), we can mine fascinating insights from PMQs debates on Prime Ministers' debating styles over the years, what topics dominated the national conversation, and even how linguistically complex different political parties' politicians are! Our speaker, Ned Stratton, will give an in-depth demo in using practical Python and Power BI to gather, prepare, and analyse text-heavy datasets and present their insights impactfully in a dashboard. As bonus material, he will also reveal "Stratton's number" - the minimum threshold of MPs in Parliament that a minor party needs to win to gain a critical mass of attention, credibility and legislative influence. Please also register for Microsoft Reactor building security access here - takes 30 seconds\, save 10 minutes on the day	Analysing PMQs with Power BI, Python and data science (almost) - Ned Stratton

Hansard online archives all debates in the UK Parliament going back over 100 years. The set-piece event of this is arguably PMQs, a cantankerous weekly 30-minute televised head-to-head between the Prime Minister and the Leader of the Opposition.

Since Tony Blair took office in 1997, PMQs debate transcripts have generated nearly 5 million words. Turning this vast bank of unstructured text data into a focused and rich analytical dataset to produce insights into trends in UK politics sounds advanced and 'data-sciency', right? Do we need sentiment scoring and word2vec? An LLM?

No, as it happens. With some subject-matter affinity and mostly vanilla Python (and some help from its text library NLTK), we can mine fascinating insights from PMQs debates on Prime Ministers' debating styles over the years, what topics dominated the national conversation, and even how linguistically complex different political parties' politicians are!

Our speaker, Ned Stratton, will give an in-depth demo in using practical Python and Power BI to gather, prepare, and analyse text-heavy datasets and present their insights impactfully in a dashboard. As bonus material, he will also reveal "Stratton's number" - the minimum threshold of MPs in Parliament that a minor party needs to win to gain a critical mass of attention, credibility and legislative influence.

Please also register for Microsoft Reactor building security access here - takes 30 seconds\, save 10 minutes on the day

Activities & events