Research Projects During Undergrad (2021 - 2022)
2021: Meta-Analytic Database for Big-Five Personality Outcomes
At the Cambridge Persoanlity & Social Dynamics Research Group, I initially worked on a large scale database of personality-outcome correlations, for which I actively improved data quality and developed a more efficient database structure.
I created statistical summaries and visualisation of the database when presenting to the group's principle investigator, Professor Jason Rentfrow.
I was also invited to contribute to the group's anonymous peer-review tasks given by a number of leading academic journals.
2021: Network Assessment for Organisations
During winter 2021/2, I took up a post working with Professor Vecchi of the Cambridge Judge Business School to develop an internal social network assessment tool.
I took on this opportunity to further my personal research interest in Social Network Analysis.
The CJBS Network Assessment Tool was proposed to provide network analyses to professionals and organisations to help increase psychological wellbeing, productivity, and resource efficiencies.
2022: Vaccine Hesitancy in Twitter Social Networks
Invited by the Cambridge Social Decision-Making Lab, I 2nd-authored the paper "Social Media Behaviour is Associated with Vaccine Hesitancy", publised by PNAS Nexus.
In this paper, I contributed to the online social network visualisation and analyses that revealed political and opinion polarisation in the US and the UK.
Main takeaway is that the US is clearly structurally polarised along politics, while it is not so clear in the UK.
You can read it here.
2022: The Power of Courageous National Culture
I completed my B.A. Dissertation co-supervised by Professor Rentfrow (Cantab), Professor Gvirtz (now KCL), and advised by Professor Götz (UBC).
We performed 20,000+ OLS predictive models on a large international dataset collected in collaboration with the TIME Magazine.
Main finding was that Courage can be used to distinctly measure national culture, and that a more courageous national culture has meaningful national outcomes such as innovation and terrorism.
The manuscript was presented at the 2022 Regional Cultural Differences Conference. Link coming soon.
2022: Adaptive Algorithm for Smart Experiments
My summer internship at the MRC Biostatistics Unit resulted in an algorithm for Smart Experiments:
Smarter A/B Tests that avoid churn from bad experiences, or adaptive clinical trials that save lives from potentially worse treatments.
What's cool about this algorithm is that it is only 10 lines of python code and can test multiple things at once.
We wrote it into a paper and it is now accepted by the 2023 Computing Conference; preprint available here.
2022: Predicting User Preferences of AI-Generated Content from Online Behavioural Patterns
Genei.io is a website that uses AI to generate summaries for the documents that users upload - it saved my life many times when I needed to read a dozen papers a day back in undergrad.
During summer, we worked together to see if we can predict users' satisfaction of an AI-generated summary from users' behavioural patterns - with limited success. I did manage to learn advanced technical skills from those at the frontier of innovation, for which I am extremely grateful.
2022: Measuring Susceptibility to Misinformation
I am fortunate to be invited by my previous supervisor, Dr Rakoen Maertens, to his big methodological work on the Misinformation Susceptibility Test (MIST).
This work is aiming to provide a framework that measures if someone is more easily influenced by misinformation or fake news. I made some minor contribution to the conceptual integration with a new factor analysis method called Exploratory Graph Analysis.
This paper is now accepted to the journal Behavioural Research Methods. Preprint available here.
Project Ideas
Here's some of my research ideas that I had to put on hold due to other workload.
I'll get back to them when I've got more time, but if you find any of these interesting please reach out for a chat! Happy to collaborate.
Modelling the Social Dynamics of Start-up Teams
Start-up cultures can range from being nice and kind (Airbnb and Yonder), nerdy and cool (Genei.io), to brutal and cruel (Facebook, apparently).
What determines a start-up's culture? Can we mathematically or computationally model it? How does it relate to start-up success?
Some methods involved might be NLP sentiment analysis and actor-based network models. Similar exploration are currently taking place at Northwestern on NASA deep-space mission crews. The challenge with start-ups is the data collection and time-span requirements.
Academic Writing Style NLP + Network Analysis
Can we algorithmically capture writing style features?
I want to see how writing style features may influence academic citation performance, how an academic's writing style may change over their career, and how certain features (e.g., "significant") might have cascaded through the academic social network.
We might also be able to see the emergence of subfields through time just by constructing time-series'd semantic networks.
Adding the Time dimension to geographic psychology
Geographic psychology studies the distribution of psychological characteristics across geographical space, such as how mental wellbeing distribute in the country. These research can help individuals find happier places to live and help policymakers to have regionally suited policies.
I want to add a time dimension to these analysis, using established methods from climate modelling to capture not just the spatial distribution but also the temporal trends of psychological variables.