Identification and estimation of causal effects in social networks, i.e. when an actions of one individual can potentially affect the outcomes of other individuals, are becoming increasingly studied in healthcare, economics, and social science. The current tools available for analysing and modelling the increasingly available data on these 'interferences' in networks are lacking. In this project, machine learning tools are being used to better understand peer effects in such networks, with a particular focus on people using microfinancing services.
This project received funding from the Turing-HSBC-ONS Economic Data Science Awards 2018.
By observing the network connections among the individuals in a social network it's possible to identify patterns of cause and effect, or 'interference', between individuals. Datasets allowing for the measurement of these network connections are also increasingly available. This project aims to use machine learning to understand these peer effects better, particularly when datasets contain missing information about certain links and connections in a network.
The example dataset that this research will use to analyse concerns the process of one person in a social network being given information about micro-financing services, and the effect this has on other people's subsequent actions. Micro-financing being financial services, such as loans, savings, insurance and fund transfers to entrepreneurs, small businesses and individuals who lack access to traditional banking services.
This work will contribute to the study of development economics (the economic aspects of the development process in low income countries), educational economics, and labour economics studies. This could help with policy makers' decisions and improve social welfare.