A cutting-edge Tropical Water Quality Hub ‘Big Data’ project has developed new ways of integrating and analysing large amounts of information from social media and other sources to give new insights into the tourist public’s perception of the Great Barrier Reef.
Dr Susanne Becken from Griffith University and her Tropical Water Quality Hub project team are using cutting-edge computing methods, including machine learning, to analyse a huge amount of posts on Twitter, Facebook and Chinese social media service Weibo sent from within a geographic ‘bounding box’ covering the Great Barrier Reef.
The system detects certain keywords relating to the poster’s perception of quality of coral, wildlife and other indicators on the Reef.
Collating and analysing the vast amount of information involved is a challenge, but since the project began in 2016 the team has improved their data-sifting capabilities, including the introduction of a new technique called ‘BoostIso’ that speeds up results by reducing the amount of duplicated calculations.
Using these techniques and others, Dr Becken and her team analysed the posts and revealed some interesting insights into the attitude of Great Barrier Reef tourists.
These social media posts form part of a global conversation on the health and future of the Great Barrier Reef, framed by the back-to-back mass bleaching events in 2016 and 2017. A significant proportion of this media coverage gave a very bleak outlook for the GBR’s current and future condition, with one outlet controversially proclaiming the GBR to ‘already be dead’.
However, despite these grim descriptors, Dr Becken’s analysis reveals that most social media posts made from within the region reflect a positive experience on the Great Barrier Reef.
“Bleaching was not mentioned by local tweets that much at all, with only 90 out of 1500 tweets containing references to it,” Dr Becken said.
“People who are there on the Reef are quite happy and they don’t seem to post about bleaching.”
“Facebook posts, with their higher wordcount, tend to start conversation threads about environmental health, while posts from Twitter are shorter and often have more negative content.”
The innovative social science project was a controversial choice for investment through the NESP Tropical Water Quality Hub, but despite only starting this year, it has already gained prominent media coverage in The Conversation and worldwide media interest.
It also links to other NESP TWQ Hub projects including one scoring aesthetic values on the Great Barrier Reef by analysing imagery posted on social media.
“This project is using over 12,000 images from Flickr, 20 per cent of which had exact location coordinates,” Dr Becken said.
“The algorithm in use here can recognisably recognise 10 species like sea turtles, coral trout and others, and we are looking to access Instagram feeds as well.”
While these immediate insights are interesting, the Big Data approach has much potential as a monitoring tool in the future.
“The way we see it, social media analysis can fall on the other end of the spectrum to a traditional professional analysis, and it can be approached as a hybrid system.
“We are also investigating the possibility of integration with the eAtlas. If stakeholders agree we could bring these systems together.
“The way we see it, analysing text posts will be good for revealing human information like attitudes toward reef conditions, while analysing images could be used for environmental monitoring.
“This approach has a wide range of possible applications in environmental monitoring, not just on the Great Barrier Reef but elsewhere.”