Journocoders: Automatically categorising things with AI

​This month we'll be learning how to use AI to give categories to our data.

​Categorisation is one of the most practically useful applications of AI for data journalism. It allows us to answer questions such as: Is this document about the topic I'm reporting on? Is this comment hate speech? Is this online review about the issue I'm investigating? Answering these questions in bulk makes certain stories possible. For example, the Washington Post reported on apps in Apple's App Store with large numbers of reviews mentioning sexual abuse. And now, with modern large language models (LLMs) it's become easier.

​In this session we'll be following a tutorial that takes us through connecting to a LLM, writing structured prompts that can classify text into predefined categories, submitting prompts in bulk, as well as evaluating and improving on the results.

​​​​​All of our events are suitable for beginners, and no programming experience is required. Bring a laptop along as this a practical, hands-on workshop. Please also sign up for a Dropbox account if you don't already have one so you can edit the shared doc we'll be using during the event.

​​​​​Schedule
7:00 🚪 Doors open
7:30 🗣 Show and tell
7:40 💻 Tutorial
9:00 🍺 Drinks at the Prince Arthur

​​​​​If you can't make the main event, you're also welcome to just join us in the pub!

​​​​​Our events do often fill up, but if it's full please do join the waitlist as spaces do typically become available. And if you find you can no longer make it, please update your registration so someone else can take your spot.