Detect clickbait with machine learning
With this one weird trick you can build a text processing pipeline!
We've all fallen for clickbait articles online. They pollute our news feeds and make it harder to filter out valuable information. In this workshop we'll stream news articles in real-time and detect clickbait using simple machine learning techniques.
By the end of the workshop you'll have your very own python app for streaming real-time news and detecting click bait. In the workshop we'll cover:
- Streaming data from a REST API
- Preprocessing textual data
- Training a simple machine learning classifier for clickbait
- Putting everything together in a scikit-learn pipeline
- Analysing our results (which news source is the most clickbaity?)
This session is an interactive workshop - please bring a laptop.
Please try to have numpy, pandas, jupyter, scikit-learn, matplotlib and flask installed in advance. The easiest way to get all these is to install Anaconda. The materials for the workshop are on GitHub (cytora/clickbait-workshop).
6pm–7.30pm, Wednesday 3rd May 2017
Southampton University, Highfield Campus, Building 6, Room 1083 (map)