Skip to main content

Tracking politicians' campaign promises on traditional and social media

Primary supervisor

Yuan-Fang Li


  • Robert Thompson

Develop NLP tools to track politicians’ campaign promises on traditional and social media: With applications to Australian, Indian and/or US politics.

This project is an opportunity for an excellent student to develop and demonstrate expertise in NLP. The project will develop tools that can track the extent to which and ways in which politicians’ campaign promises are disseminated through traditional and social media during election campaigns. For selected Australian, Indian and/or US election campaigns, we will identify the campaign promises of each of the main parties and candidates from their election manifestos, platforms and official websites. The NLP tools to be developed and applied in this project will then track these campaign promises in traditional and social media. The analysis of traditional media will focus on digitally archived national and regional newspapers, while the analysis of social media will focus on Twitter data. The NLP tools to be deployed in this project will identify the frequency and contexts in which different campaign promises receive attention. These tools will also quantify the ways in which campaign promises are featured, for instance whether the promises are referred to positively or negatively by candidates and citizens.

This project is of significant academic and practical relevance, and we expect it to lead to one or more high-profile international publications on which the student has the opportunity to become a co-author. In terms of NLP research, the project will advance supervised text classification methods, including sentiment analysis, word embeddings and neural networks.

The project is also positioned at the cutting edge of comparative Political Science. It is widely acknowledged that for democracies to function effectively, political candidates must offer voters meaningful choices during election campaigns. This involves parties making campaign promises on policy issues, which are communicated to voters through traditional and social media. This project therefore examines processes that lie are at the heart of democratic practice. The elections that may be the focus of the applications include the 2019 Australian national election, the 2019 Indian general election and the 2016 and/or 2020 US Presidential elections.

Required skills

A foundation in NLP, supervised text classification methods, including sentiment analysis, word embeddings and neural networks. An affinity with the politics of India and/or the US.

The supervisory team

The student will be supervised by Yuan-Fang Li in the Faculty of IT.
Prof. Robert Thomson in the School of Social Sciences will provide supervision on the Political Science aspects of the project.

International collaborations

The student will become part of an interdisciplinary and international network of prominent researchers in this field. The project is part of the University of California San Diego-Monash joint project "AI for Stronger Democracy and Policy Performance", which was initiated by Mark Andrejevic (Monash School of Communications), Wray Buntine (Monash Faculty of IT), Seth Hill (UCSD, Political Science), Ndapa Nagashole (UCSD, Computer Science), Christina Schneider (UCSD, Political Science), and Robert Thomson (Monash, Social Sciences).

The project is also linked to an international network of Social Scientists and Data Scientists led by Profs. Elin Naurin (University of Gothenburg) and Robert Thomson (Monash), funded by the Bank of Sweden. This network is developing both qualitative and quantitative automated methods for analysing what political parties promise to voters during election campaigns:

In addition, the project is linked to an ongoing established Political Science project on campaign promises: the Comparative Party Pledges Project, which is co-led by Professors Naurin (Gothenburg), Royed (Alabama) and Thomson (Monash). A summary of some of the work of this project can be found here:






Student cohort

Single Semester
Double Semester