ProfessorTristram Alexander

Project Opportunities

Title: Language dynamics on social media

Summary of opportunity:

Development of a language evolution model using social media data.

Opportunity synopsis:

What is happening on social media? Who is speaking to whom? The language of social media is rapidly evolving, but we don’t know how or why. The enormous quantities of data available from Twitter allow for a real-time investigation of the exchanges occurring on this platform. By focusing on specific elements of this data, such as hashtags, this project will quantify the rate of evolution, identify the drivers of this evolution, and develop a model for the language dynamics. This project will use information theory and techniques from machine learning and Natural Language Processing for the quantitative analysis.

--------------------------------------------------

Project Opportunities

Title: Fighting the spread of misinformation

Summary of opportunity:

The development of a quantitative measure of the spread of misinformation, for the purposes of developing a strategy to counter science denial.

Opportunity synopsis:

Why aren't scientific results having more impact on public opinion, policy and political will? The use of misinformation by vested interests has a long history, however the advent of social media, and the fracturing of the media landscape, has vastly increased the power wielded by those bent on introducing confusion into debate. The sophistication of the deception, coupled with the severity of the consequences, has prompted a large amount of recent research into the nature of misinformation [1-3], and how best to combat it [4,5]. High impact research has recently been published on the spread of `fake news' [1-4], and the need to combat scientific misinformation [5,6], however there has been no attempt to quantify the spread and impact of scientific information on public opinion and policy development. 

While social media brings unprecedented reach, it also allows for accountability, as the connections between members of the network are visible. This project will use the tools of network analysis to quantify the nature of information flow within, and between, the networks of scientists, the general public, and key influencers (such as politicians and media organisations). This quantitative analysis will give a measure of the extent to which influencers are relying on misinformation versus science in the development of policy. The project will also identify the characteristics of ``super-spreaders" in the network, and seek to determine how these characteristics may be harnessed for a public innoculation strategy. Ultimately this project is focused on tackling the fundamental barriers to the use of science in society, by developing quantitative measures to first diagnose the problem, and then developing a strategy to overcome the barriers. 

[1] N. Grinberg et al., Science 363, 374 (2019). [2] A. Bovet et al., Nat. Comm. 10, 7 (2019). [3] S. Vosoughi, et al., Science 359, 1146 (2018). [4] G. Pennycook et al., Proc. Natl. Acad. Sci. U.S.A. 116, 2521 (2019). [5] J. Farrell et al., Nat. Clim. Change 9, 191 (2019). [6] M. Buchanan, Nat. Phys. 15, 731 (2019).

--------------------------------------------------

Project Opportunities

Title: Data-driven modelling of social systems

Summary of opportunity:

The tools of physics will be applied to social dynamics data, to develop a predictive model, and ultimately inform policy decisions in the chosen area.

Opportunity synopsis:

The availability of immense quantities of data have made it possible to apply the techniques of physics to complex social problems, such as gun violence and political polarization. There are a wide range of possible research directions contained within this umbrella title, however all projects are united in the underlying methods being used. Data is used to develop a simplified model of the scenario in question, this simplified model is then used to make predictions, and these predictions are compared with the physical reality. A significant part of the project will be the development of suitable metrics to render the social phenomena of interest into a quantitative form. The specific tools of physics will then depend on the scenario. For gun violence for instance, agent-based modelling coupled with statistical mechanics and gun violence data can be used to explore different gun control measures, while political polarization can be explored using social media data, topic modelling and neural dynamics models. For more details on specific projects within this category, please contact the project supervisor.