Data Science
HARNESSING DATA FOR CHANGE
The Shaftesbury Partnership uses Data Science in three main areas: Simulation, Measurement, and Prediction.
Simulation: Decoding Complex Challenges
We use agent-based modelling to deeply understand today's complex challenges.
Together with our partners, we harness this approach to best understand how inequality and other aspects of broken systems can be simulated and modelled. Visit our Inequality Simulator page to explore an early prototype under development.
Our approach is to focus on particular sectors such as healthcare, migration, and the climate, and model the systemic interactions between agents using continuously improving agent-based simulations.
This information can then be used to inform decision makers (in combination with the use of war gaming and red-teaming decisions processes), through simulating interventions in these and other models.
Measurement: Elevating Intervention Effectiveness
Randomised control trials are our gold standard for measuring effectiveness.
We are interested in taking this to the next level through working with our partners using mobile and other sensing technologies to bring down cost, generate more real time data, and make the practice of such approaches mainstream.
Bringing this data into contexts beyond healthcare, and other sectors such as education and offending, with tools to speed up the process of measurement will enable greater insights to be brought into practitioner contexts beyond pure academia.
Our approach promotes quick decision-making and innovation, countering delays often seen in regulated sectors.
Prediction: AI-Enhanced Decision Making for Tomorrow
Together with our partners, we use AI to improve predictions and decision-making.
Whilst AI is very powerful it can often be limited to predicting the future based on past datasets which are not always great at dealing with future uncertainty, and which can at times exacerbate inequalities.
We focus on AI technologies that enhance group decision-making, leading to improved policies and actions.
The tools our partners are developing have wide application in boardrooms, government, and communities, and lead to greater levels of satisfaction in the process than other forms of decision making including voting.