Friday lunchtime lecture: Improving UK government spending using data science

In February 2015 the UK government launched Contracts Finder, a portal that advertises open government-spending contract opportunities, allowing companies to easily and quickly bid for them. The UK spends £250 billion every year on public spending contracts, including vital public services like the NHS and it’s important that this money is spent as efficiently as possible. As part of an ASI fellowship, William Jones worked in collaboration with the UK Cabinet Office to increase the number of companies bidding on these contracts. He did this by automatically recommending them contracts which are available and relevant, using publicly accessible information from UK companies. The web application William built is called Contracts Recommender. His talk will explain the public sources of information he uses as well as how Contracts Recommender works. William Jones was recently awarded the MPhil in Advanced Computer Science from the Cambridge Computer Laboratory with Distinction. During the ASI fellowship he first worked on a deep learning image classification project with Tractable.io, using the convolutional neural network framework Caffe. Following this he joined a project with the Cabinet Office using natural language processing to automatically recommend relevant government procurement contracts to companies in the UK. He has just started a PhD in Mathematical Genomics at Cambridge.