Tropical Water Quality Hub researchers are employing Big Data and artificial intelligence approaches to develop and test methods of analysing thousands of social media posts with the potential to turn every visitor to the Great Barrier Reef into a citizen scientist. The project team for NESP TWQ Hub Project 5.5, including Professor Bela Stantic and colleagues at Griffith University, applied sorting models based on deep learning techniques to a human-annotated database of over 13,000 human-annotated Tweets posted within the Great Barrier Reef area. The tweets were then allocated into 11 different ‘labels’ or topics of conversation such as the Great Barrier Reef, travel, etc. and tweets were also categorised as being of either ‘positive’ or ‘negative’ sentiment. The project confirmed a high level of accuracy and it is hoped that advanced techniques like these will enable researchers to leverage the social media posts of visitors to the Great Barrier Reef and other important areas – a huge and inexpensive data set of observations, attitudes and intents that could be highly useful information to decision-makers.
Photo: Griffith University