The Essex Centre for Data Analytics combines data with the experience of people and communities across Essex in order to create powerful insight and inform policy decisions.
The Data in ecda is much more than a large collection of numbers arranged in vectors and matrices. Data for insight necessarily requires collection, storage, transformation, aggregation, and most importantly learning. Traditionally, learning from data —also known as statistical learning— uses statistical models to estimate a mathematical function for purposes of inference and prediction. In order to use statistical learning and create powerful insight, analysts must have a command of a vast menu of statistical tools and an excellent theoretical understanding of the problem at hand.
In this light, ecda relies on the constant training of its own People: analysts, program managers, communication officers, and partners across organizations. The partnership between ECC, Essex Police, and the University of Essex plays a key role in this effort. Essex Partners work together to design the content of ecda learning modules, which are delivered by the University’s world-class researchers based at the Institute for Analytics and Data Science and the ESRC-funded Business and Local Government Data Research Centre, among others.
As a Reader at the University of Essex and Chief Scientific Adviser to ECC, it is a privilege to work with colleagues from ECC and Essex Police to map the current set of skills and identify opportunities for further training on statistical learning. For instance, in one module we may explore hypotheses testing and the linear model, while in another we may investigate methods for dimension reduction or tree-based statistical models. We implement these models in the R Project for Statistical Computing; this software for statistical computing is free and powerful, and it has useful graphical tools for data visualization. Colleagues at ECC and Essex Police are already familiar with R and we hope to contribute to a more extensive use of this computing environment across the Partnership.
As ecda continues to mature, we will strengthen the skill set of People analysing the large amounts of data that ecda will deliver. I anticipate further training on statistical foundations as well as more advanced modules on data science and scientific research in general. This may include more models in supervised learning, but also on unsupervised learning for the detection of structure in unlabelled data.
We want to get things right and generate valid insight from Data. Therefore, I am interested in complementing our training with modules on resampling methods (for instance, cross-validation) and robustness tests. I am also an advocate of a strong theoretical understanding of policy challenges and training on causal inference and experimental methods would make a great addition to our current set of skills at ecda. This will allow us to generate insight while understanding the processes that drive key challenges such economic inequality or domestic abuse, among others. In this light, we will continue to invest in Data, but also in the People that collect, process, and analyse it. This is crucial for our aspiration to make Essex an international leader in the use of data science and artificial intelligence for public policy.