Edge Hill University and Career Connect KTP 24_25 R3

A machine learning system will scan local government records and social media posts to predict which young people are at risk of dropping out of education, employment, or training. The problem is that current methods for identifying NEET (Not in Education, Employment, or Training) young people rely on crude demographic categories that miss individuals whose risk is tied to their specific local circumstances—poor transport links, a declining local industry, or family instability. This project combines natural language processing with place-based data to build risk profiles that account for both personal history and neighbourhood conditions. If the approach works, local authorities could intervene earlier and more precisely—offering targeted support to a teenager in a specific postcode rather than blanket programmes for an entire region. The system would quietly improve the way public services triage resources, without requiring young people to self-identify as at risk. This is applied research with a clear practical endpoint: fewer young people falling through the cracks between school and work.