Development of Interactive Visualisation Software for Research Communication

Published: May 13, 2021

By: J. Boshoff 1, I.V. Cristiani 1, T. Sibanda 2, M. Kolakshyapati 2, D. Schneider 3, M. Welch 3 and I. Ruhnke 2 / 1 CASI Data Transformation Hub, University of New England, Armidale, NSW, Australia; 2 School of ERS, Faculty of SABL University of New England, Armidale, NSW, Australia; 3 School of Science and Technology, Precision Agriculture Research Group, UNE, Armidale, NSW, Australia.

Tracking and understanding hen movement in commercial free-range flocks is challenging due to complex housing furniture as well as the sheer volume of individual data involved. The purpose of this study was to develop interactive visualisation software that allows users to display recorded data and individualise graphs on demand using existing data. A total of ~9,375 laying hens were housed in three identical commercial free range sheds equipped with an open aviary system. Hens were individually tracked within the sheds as well as on the range from 16 - 72 weeks of age using a custom made RFID system. Hen movements were captured every second generating >1.597 billion records (~1 TB). Data were consolidated in PostgreSQL [v10.5, PostgreSQL Global Development Group] an open source relational database. “Materialised views” were built to summarise information and format data for visualization. Web based software was build using Javascript [v 1.8.5, Netscape Communications, Dulles, US] for both front and back end development. AngularJS [v1.6, Google LLC, Mountain View, US] served as front-end web application framework. D3.js [v5.5.0, M. Bostock & J. Heer, New York, US] was used to render visualisations. We developed interactive web based software (visualhens.une.edu.au) enabling users to visualise the movement and visitation patterns. Users can generate visualisations of:

Percentage of hens selected that were detected at different locations;
Origin, destinations and numbers of movements of groups of hens in the aviary system;
Movements of individual hens to determine if some hens remain in certain locations.

AUSTRALIA - DEVELOPMENT OF INTERACTIVE VISUALISATION SOFTWARE FOR RESEARCH COMMUNICATION - Image 1

Hens can be selected by weight as well as by activity classifications. Temporal selections enable users to visualise specific days and events.

In conclusion, our dataset and web tool can help answer specific questions on range and structure utilisation. Open source tools and relational databases can be harnessed to analyse and visualise large datasets. In future, additional datasets e.g. weather and necropsy data can be added allowing for additional features.

ACKNOWLEDGEMENTS: This research was funded by the Australian Eggs.

Abstract presented at the 30th Annual Australian Poultry Science Symposium 2019. For information on the latest edition and future events, check out https://www.apss2021.com.au/.