Senior Departmental Research Lecturer in Social Data Science
Charles is a social science methodologist and applied social data scientist with a background in high-dimensional econometrics, having completed his PhD in 2016. He is both a Senior Departmental Research Lecturer at the Leverhulme Centre for Demographic Science, and an Associate Member of Nuffield College. He also sits on the Steering Group of Reproducible Research Oxford, and is a Co-Investigator at the ESRC Centre for Care. As part of his lecturing, he co-convenes Demographic Analysis, Life Course Research, and the Oxford Partner site of the Summer Institute in Computational Social Sciences. In the past, he's taught modules related to financial econometrics, Python for sociologists and statistical software more generally, and replication in open social science. He has recently given workshops and guest lectures on the themes of 'An Introduction to Machine Learning', 'An Introduction to the Command Line' and `LaTeX in 105 Minutes'. He is always interested in hearing from potential co-authors or prospective graduate students who share his enthusiasm for using Python, R, and LaTeX.
Charles is particularly interested in unique Big Data origination processes -- be they unstructured or otherwise -- and how they contribute to social inequality, mobility and stratification more generally. Other current areas of interest include the machine learning methods, civic technology, spatial and time series econometrics, model uncertainty, and scientometrics.
Specific projects underway at present include:
A large scientometric review of the 'Evolution of Science';
Interperable ways to directly compare predictive systems (the 'Inter-Model Vigorish');
Work on inequalities in life expectancy across the very long run (the 'Legacy of Longevity');
A library to incorporate model selection, averaging and influence analysis into robustness pipelines;
Full list of publications Google Scholar, and more information on his academic homepage.
Follow the development of his programming on GitHub.