Our ‘Data Science Student of the Year’ award highlights data science research projects conducted by students, as well as the relevance and impact of their research work. Revisit the projects from this year’s DatSci Awards by clicking the thumbnail images next to each finalist’s name.

Winner: Rory Boyle 🏆
The Whelan Lab – Trinity College Dublin

This project developed an accurate and interpretable model of brain-predicted age differences (brainPAD) using open-access MRI data. The model was validated on external datasets, containing rich cognitive data. The brainPAD cognitive function relationship was assessed in each external dataset in order to investigate its utility for objective measurement of cognitive function, which is critical in order to identify individuals at risk of significant cognitive decline.

Finalist: Gaurav Pahuja
SSE Airtricity

Data Optimisation Network is an end to end application which connects multiple databases together to create a 360 degree view for the business on a same geospatial index through using machine intelligence and data science techniques. DON will seek to accelerate the growth of the business. It will work as a proactive and reactive solution platform which will enhance business strategic decisions.

Finalist: Amal Saadallah
TU Dortmund

Matching supply with demand in dynamic environments is one of the biggest issues faced by many industries such as the taxi industry. BRIGHT: a drift-aware supervised learning framework is proposed to predict short-term horizon taxi demand through a creative ensemble of time series analysis methods that is able to handle distinct types of changes/concept drifts to cope with the dynamic behaviour of urban mobility patterns.

Finalist: Andrew Kenny
University of Limerick

Opinion polls are becoming less reliable in predicting political views, as seen in high-profile cases such as the 2016 US Presidential election and Brexit. This research proposes that the application of machine learning to social media can provide a viable alternative, while also addressing underlying issues found in traditional methods such as social desirability bias and confirmation bias.

Finalist: Dixon Vimalajeewa
Waterford Institute of Technology

The social network analysis (SNA) is now commonly used to characterise the behavioural dynamics of social groups. The increasing complexity of SNA data necessitates the investigation of novel strategies to transform such data into useful metrics, which can subsequently be used to support day-to-day decision making. A novel matrix, Animal importance is derived based on GPS mobility data to explore and identify atypical social behaviours in cows in smart dairy farming.

Finalist: Suad Al Darra
National University of Ireland, Galway

Are refugees a threat? How do news articles describe refugees and migrants? This work highlights the issue of xenophobia against displaced people by applying the latest data science methodologies over open-source news datasets to train a classifier that could automatically detect xenophobia language, and also to find a better visual way to tell the narrative about refugees.