Using data science to revolutionise evidence-based practice for environmental management

Project supervisors

Dr Carly Cook, Faculty of Science (Main Supervisor)
Dr Jessica Walsh, Faculty of Science
A/Prof Peter Bragge, Monash Sustainable Development Institute
Dr Brea Kunstler, Monash Sustainable Development Institute

PhD project abstract

Environmental managers must make decisions about the most effective actions to protect biodiversity. Ideally, these decisions would be made on the basis of high-quality evidence, enabling managers to weigh up information about the costs, benefits and probability of success of alternative approaches to management. Yet managers rarely have access to the best available scientific evidence to support their decisions, leaving them to make important decisions that may do more harm than good.

One approach to supporting managers to making evidence-based decisions is to identify and synthesise the available evidence and provide it to them in a format they can use to guide their decisions. However, this type of evidence synthesis requires exhaustive searches of the available literature, filtering evidence by relevance and quality and then integrating the findings into actionable recommendations for practice and policy.

Traditional approaches to this type of evidence synthesis, drawn from the field of evidence-based medicine, involve reviewing primary evidence studies addressing a focused question. These can be extremely time consuming, taking more than a year in many cases. Over the last 10 - 15 years, ‘rapid reviews’ or ‘reviews of reviews’ have gained traction in addressing broader questions in less detail and enabling actionable recommendations for policy and practice to be developed in weeks or even days. This has coincided with growth in the review sciences beyond medicine to a broad range of disciplines including environmental management, education, business and the social sciences. There is also significant interest in the potential for tools from data sciences and artificial intelligence to revolutionise evidence review processes. There has been some progress using artificial intelligence to support evidence synthesis, and web-based ‘evidence maps’ enable visualisation of the strength of  evidence for interventions and outcomes in a defined domain.

The conservation literature poses a unique challenge because of the broad range of species and interventions, the lack of standardised terminology and large amounts of evidence outside the peer-review literature. A recent review funded through the Monash Interdisciplinary Research Support Scheme (“Combining rapid evidence synthesis and expertise for threatened ecosystem recovery”, Lead Dr Jessica Walsh) applied rapid review and evidence mapping to the challenge of woodland management. The resulting evidence map was well-received by environmental managers, indicating that these techniques have potential to bridge the evidence-practice gap in conversation.

This project will build upon this initial work to identify, design and adapt tools from data science to support all stages of evidence synthesis with the goal of improving conservation decisions.

Areas of research

Evidence-based policy and practice, AI & Data science, Better Governance and Policy

Project description

Environmental managers must make decisions about the most effective actions to protect biodiversity. Ideally, these decisions would be made on the basis of high-quality evidence, enabling managers to weigh up information about the costs, benefits and probability of success of alternative approaches to management. Yet managers rarely have access to the best available scientific evidence to support their decisions, leaving them to make important decisions that may do more harm than good.

Over the last 10 - 15 years, ‘rapid reviews’ or ‘reviews of reviews’ have gained traction in addressing broader questions in less detail and enabling actionable recommendations for policy and practice to be developed in weeks or even days. This has coincided with growth in the review sciences beyond medicine to a broad range of disciplines including environmental management, education, business and the social sciences. There is also significant interest in the potential for tools from data sciences and artificial intelligence to revolutionise evidence review processes.

This PhD project aims to address the challenge of optimising environmental management by applying rapid evidence review and data sciences to address this need. A recently completed project has shown that environmental evidence in a defined area (conversation management for woodlands) can be rapidly reviewed and made practically applicable through an evidence map. The PhD will involve:

  • Identification of an environmental challenge / topic area where there is a need for evidence resources to inform management - the team has access to policymakers and other environmental experts who can articulate priority areas for consideration
  • Formation of a community of practice that brings together policymakers, environmental experts, review scientists and IT / data scientists
  • Registration and completion of a systematic rapid review addressing a high-priority question
  • Identification and testing of relevant tools from data science to assist in the different stage of evidence synthesis
  • Development of an online evidence map representing identified interventions against relevant outcomes, which will highlight areas of strong, weak and non-existent evidence
  • Exploring the usefulness of the evidence map from the perspective of policymakers, environmental managers and other stakeholders
  • Developing and testing approaches to keeping the evidence map updated as a ‘living’ resource through periodic application of the review protocol to capture emerging evidence at defined time points (e.g. every six months)

PhD student role description

Are you passionate about environmental management and interested in how research and data science can inform environmental managers about what works? This PhD offers a unique opportunity to address the challenge of using data science to streamline the identification and application of evidence-informed practices to environmental management.

Environmental managers must make decisions about the most effective actions to protect biodiversity. Ideally, these decisions would be made on the basis of high-quality evidence, enabling managers to weigh up information about the costs, benefits and probability of success of alternative approaches to management. Yet managers rarely have access to the best available scientific evidence to support their decisions, leaving them to make important decisions that may do more harm than good.

Over the last 10 - 15 years, ‘rapid reviews’ or ‘reviews of reviews’ have gained traction in addressing broader questions in less detail and enabling actionable recommendations for policy and practice to be developed in weeks or even days. This has coincided with growth in the review sciences beyond medicine to a broad range of disciplines including environmental management, education, business and the social sciences. There is also significant interest in the potential for tools from data sciences and artificial intelligence to revolutionise evidence review processes.

Working with a team of evidence review experts, environmental scientists and AI / data scientists, your PhD will involve:

  • Identification of an environmental challenge / topic area where there is a need for evidence resources to inform management
  • Formation of a community of practice that brings together policymakers, environmental experts, review scientists and IT / data scientists
  • Registration and completion of a systematic rapid review addressing a high-priority question to inform best practice
  • Identification and testing of relevant tools from data science to assist in the different stage of evidence synthesis
  • Development of an online evidence map representing identified interventions against relevant outcomes, which will highlight areas of strong, weak and non-existent evidence
  • Exploring the usefulness of the evidence map from the perspective of policymakers, environmental managers and other stakeholders
  • Developing and testing approaches to keeping the evidence map updated as a ‘living’ resource through periodic application of the review protocol to capture emerging evidence at defined time points (e.g. every six months)

There is strong demand for rapid access to knowledge that moves beyond long academic reports to usable, updatable databases that make sense to policymakers and practitioners on the ground. The skills developed in this PhD will position the candidate strongly in research, policy and practice arenas and enable access to a deep network across these sectors.

With numerous environmental challenges faced globally in the coming decades, the need for evidence-informed policy and practice has never been stronger. The techniques learned are also readily transferable to other environmental challenges or reviews and evidence maps in other domains. This PhD therefore unlocks a number of exciting and ongoing opportunities in environmental management, review sciences and / or AI and data sciences.

Required skills and experience

A strong background in data science, including an understanding of machine learning and software development. Strong creative and analytical skills. An interest in biodiversity conservation and applied research. Desirable - working knowledge of key review tasks such as literature searching, screening, quality appraisal and data extraction

Potential start date

Late 2021 / Early 2022

Submit your Expression of Interest