Title: Big-data credit scoring: risk management in Chinese social credit programmes
28 October 2019
4.00-5.30pm, University of Warwick, E2.02 (Social Sciences)
Ruowen examines the organisation process by which Big Data credit scoring models are produced, investigating the analytical work of data scientists who continuously maintain and improve their models to keep the results predictive. Big Data algorithmic technology is having a profound impact on our social, organisational, and public life and it permits large tech companies to perform analytics for consumer credit-risk assessments and to determine credit risk. Because of the potential for wide applicability of the scores, it is crucial to understand how big tech companies organise data analytics and verify the results before deployment.
Based on ethnographic fieldwork in a credit score modelling team of a large tech company, Ruowen's research studies the development of an emerging Big Data algorithmic credit-scoring technology alongside the government’s programme for building a social credit system in China. Her findings show that data scientist work is a continuous, repetitive, and a pre-prescribed process of developing and updating models that are complemented with machine learning-generated results, and that the way that data scientist work is organised has a direct impact on the produced model. Following this process, the model development team sees that their breakthrough of data analytical work lies in getting more data from internal and external sources to find predicting features. In parallel with the notion behind “data as an asset,” internal and external data barriers get in the way of allowing more data to enhance model predictability. This research expands the perimeter of how we look at algorithms and, broadly, other data-driven computing devices by looking at the organisational setting through which they are produced.
You can find a recording of this event here.
Ruowen Xu is a PhD Candidate in Business and Management at the Warwick Business School. Her research seeks to understand how big data systems and calculative devices are evolving and changing people's everyday life. She studies the impact of credit rating systems and more specifically, how a new form of modelling work in big data and machine learning technology is being adapted to form credit scoring algorithms. Drawing on inter-disciplinary insights, combining sociology and organizational theories, she investigates the social construction of big data and more broadly AI technology.