Peter Marendy, Senior Research Projects Officer
Thi Thuong Huyen Nguyen, OCE Postdoctoral Fellow
Paulo A. de Souza Junior, OCE Science Leader (affiliate member)
Ray Williams, Visiting Scientist (affiliate member)
Selim Mahbub, OCE Postdoctoral Fellow (affiliate member)
Ferry Susanto, PhD Student (affiliate member)
2014 – ongoing
Honey bee health is globally under threat as a result of various factors, including, agricultural intensification, pests such as the Varroa mite, the use of pesticides, and bee management practices. Colony Collapse Disorder (CCD) refers to the related decline of bee population or the entire collapse of bee colonies as a result of these and potentially other factors. A global effort lead by CSIRO, the Global Initiative for Honey Bee Health (GIHH), is being implemented with the aim to protect honey bee health, ensure sustainable production of honey bee pollination dependent crops, and increase productivity through better management of crop pollination. As part of this, a central data repository hosted by CSIRO will facilitate management of data collected on a global scale and serve as a knowledge base to learn about the honey bee behaviour as well as the potential causes underlying CCD.
Towards these goals, swarm sensing is implemented as an instrumental technology to learn about the behaviour of the honey bees as well as their responses to stressors that may affect their health. In an initial study on Tasmania, thousands of high tech micro sensors have been successfully fitted to honey bees to measure their presence and absence from the bee hives based on RFID technology. This system is able to log entrance and exit of all honey bees that are equipped with the sensor backpack and therefore allows for exact status reports on the presence of the bees in the hive. A new iteration of the sensor technology will go well beyond just measuring bee presence but will also provide information on temperature, humidity and light. The abundance of data being collected by the thousands of bee-attached sensors as well as additional environmental sensors poses a number of challenges with regard to the interpretation and comprehension of the data, both computationally as well as from a user perspective. Identifying patterns in the data to facilitate learning about bee behaviour from the recorded data is particularly important for informed decision making about the risk of CCD based on, for instance, the factors outlined above. Effectively communicating the recorded data to a variety of end users, including, honey bee farmers, scientists, and commercial partners, is crucial to convey the most relevant information in a meaningful way and in any given context.
Advanced visualisation is widely recognised as a means of communicating the complexity of science, engineering and technology to domain experts and the wider public. This project aims to be an integral part of and contributor to the GIHH by addressing the above challenges through researching and developing advanced visualisation techniques. The project focuses on visual analytics to assist a variety of end users with data comprehension, as well as with informed decision making. A flowchart of the integral parts of the visualisation pipeline is shown below. The type and sophistication of the visualisation is strongly dependent on the query by the end user and the given context. It may range from simple 2D presentation of statistics on a screen to contemporary visualisations of spatio-temporal data presented in 3D space and overlayed on natural environments. Statistical analysis and machine learning techniques are integral in preparing the data for visualisation. Appropriate user interfaces facilitate effective presentation of the visualisations on a given display and may include simple dashboards for statistical data presentation as well as augmented reality (AR) systems based on contemporary devices such as smart glasses.
The focus of this project is on visual analytics, hence, the aim is not only to develop sophisticated visualisations of honey bee behaviour and health, as well as information on key individuals (e.g., drones, queens, foraging bees), but to create a visualisation pipeline that takes into account both the end user’s desires and expectations as well as appropriate analysis techniques to facilitate the visualisations of both bee activity and environmental conditions. We believe that this user centred design approach will benefit all stake holders of the GIHH, as the abundance of sensor data is made accessible to them in a meaningful way.