movenetapp

The movenet app provides a graphical user interface to the movenet package, in the form of a Shiny app.

The goal of movenet and the movenet app is to simplify the effective use of livestock movement data in veterinary public health. It facilitates the dataflow from livestock movement data to social network analysis and disease transmission models, while addressing common data issues such as the diversity of data formats and privacy preservation.

Flow chart showing the workflows of the movenet package, as well as how movenet addresses some common data challenges.

movenet and the movenet app are being developed in the context of the NordForsk Digitalisation of livestock data to improve veterinary public health (DigiVet) project.

Disclaimer: The movenet app is under active development. The interface may change without prior warning.

Workflows

movenet and the movenet app include a range of workflows for the processing and analysis of livestock movement data and (optional) holding data:

• standardisation of data into a single format, allowing for interoperability and integration of data from different countries or systems

• pseudonymisation of livestock movement and/or holding data, to improve the potential for data sharing and collaborative analysis

• generation of network representations, and social network analysis, of livestock movement data

• integration of livestock movement and/or holding data into transmission models

• exploration of the effects of different pseudonymisation strategies on network properties, so as to allow users to find a suitable balance between the identifiability of the data and the accuracy of these properties.

For more detail on each of these workflows, see the vignettes on the movenet package website.

Installation

You can install the development version of the movenet app from GitHub.

Before you install the movenet app, you first need to install the latest version of movenet, which contains the basic functions behind the movenet app.

The process is as follows:

# install.packages("devtools")
devtools::install_github("digivet-consortium/movenet")
devtools::install_github("digivet-consortium/movenetapp")

Usage

To launch the movenet app, run the following:

movenetapp::runMovenetApp()

Example use case scenario 1

You work at a veterinary health agency, and want to collaborate with external experts to model the spread of a potential ASF outbreak in your country. You need to make livestock transport data non-identifiable, but you are concerned about the impact that modifying movement dates may have on any analyses.

1. Input and reformat your movement dataset
2. Apply a range of different date modifications
3. (View the datasets to check the results)
4. Generate networks for some or all modified datasets
5. Calculate network measures across your networks (currently implemented for maximum reachability only)
6. View the network measure comparison page. Decide on a jitter range or rounding unit that provides a good balance between preservation of privacy, and low impact on network measures of interest (e.g. maximum reachability).
7. Pseudonymise the chosen modified dataset. Download the dataset and send it to your collaborators. Download the key and save it somewhere safe, to de-pseudonymise any future modelling results.

Example use case scenario 2

You have just read a paper that applies network analysis to a pig transport network in another country, and discusses its implications for disease control. You are keen to know how transferable the results and implications would be to your own country’s livestock movement network, in case of an ASF outbreak. You have access to livestock movement data, and would have liked to do a similar study – but you lack the resources/time/R skills. Luckily, a quick network exploration with the movenet app could provide some clues.

1. Input and reformat your movement dataset
2. Generate a network for the movement dataset
3. Explore the data summaries and network analyses on the “Explore a single network” page. How do these compare to the results in the paper? Does anything of interest jump out at you, that would be worth looking into in more detail in a more elaborate study?