PORTFOLIO

Abstract:

The Smart Earth Project supports the mobilization of digital technologies to address some of the most pressing challenges of the Anthropocene, from climate change to biodiversity loss. Led by Geography Professor Karen Bakker, our research team studies the impacts of digital technologies on environmental sustainability. Our Smart Earth platform provides a catalogue of environment-oriented technological innovations: from citizen science apps to drones, virtual reality, artificial intelligence and many more, our database shows a real-time index of how technology is being used to model, monitor and regulate ecosystems across the globe.

Between 2020 and 2022 I designed and created an automated system for the development of the 'Smart Earth Technologies' database. The Smart Earth Technology Database is a collection of news and website articles discussing innovative digital technologies with beneficial environmental applications. Smart Earth tech innovations may be defined as digital technologies that combine: (i) digital instrumentation (e.g. sensors, satellites, drones, digital cameras); (ii) interconnectedness (e.g. IoT/ICT, wireless sensor networks, radio tags); (iii) intelligence (e.g. machine learning, computer vision, artificial life, robotics); and (iv) immersive (e.g. VR/AR). All articles in the database are classified according to a set of 10 Technology Labels and 10 Application Labels, and must belong to at least of each label category. These classes are non-exclusive meaning that each article may belong to more than one Technology Label or Application Label.

Development of the Smart Earth Database involves a combination of web scraping and machine learning-based text classification using natural language processing (NLP). I developed a series of web scraping programs to extract text contents and metadata of web articles from a variety of environmental and technology themed web series. Once extracted and structured, articles were then evaluated for ‘Smart Earth’ relevancy and classified based on the aforementioned Technology Labels and Application Labels. Initial development of the Smart Earth Database involved manual human classification on the part of the Smart Earth Team. Once this database reached a sufficient starting size of, I designed an automated classification process using machine learning. Final outputs resulting from the automated web scraping and text classification programs were then validated and a selection of articles uploaded to the Smart Earth Website.

The contents of this database would go on to stem the writing of several global award winning books (Sounds of Life, 2022) and publications (Conservation Acoustics, 2021; Smart Oceans, 2022; Digital Waters, 2022) by Professor Karen Bakker.

Project Results: