ASAS-NANP SYMPOSIUM: prospects for interactive and dynamic graphics in the era of data-rich animal science1

Abstract

Statistical graphics, and data visualization, play an essential but under-utilized, role for data analysis in animal science, and also to visually illustrate the concepts, ideas, or outputs of research and in curricula. The recent rise in web technologies and ubiquitous availability of web browsers enables easier sharing of interactive and dynamic graphics. Interactivity and dynamic feedback enhance human-computer interaction and data exploration. Web applications such as decision support systems coupled with multimedia tools synergize with interactive and dynamic graphics. However, the importance of graphics for effectively communicating data, understanding data uncertainty, and the state of the field of interactive and dynamic graphics is underappreciated in animal science. To address this gap, we describe the current state of graphical methodology and technology that might be more broadly adopted. This includes an explanation of a conceptual framework for effective graphics construction. The ideas and technology are illustrated using publicly available animal datasets. We foresee that many new types of big and complex data being generated in precision livestock farming create exciting opportunities for applying interactive and dynamic graphics to improve data analysis and make data-supported decisions.

Keywords: dynamic graphic; human–computer interaction; image; interactive graphic; statistical graphic; visualization.

Figure 1.

Plot A: the line graph of body mass in grams over age in days for each offspring with line colored by the treatment group and plot facetted by sex of the offspring. Plot B: The points and line graph in the above figure show the average body mass over age by the sex of offspring. The vertical lines through the points show the standard error of the mean.

Figure 2.

The black points and lines show the body mass in grams over age of the offspring of the quails. Each curve corresponds to one quail offspring. The gray curves correspond to the observations of all quail offspring disregarding the treatment group or the sex. The growth curve is facetted by the treatment group and sex of the offspring.

Figure 3.

The black lines and points show the body mass over time of a sample of 8 offspring, one each from the combinations of treatment group and sex. The superimposed red line is the predicted growth curve from the generalized additive model. The number in the background of each facet is the egg ID of the quail.

Figure 4.

One of these plots is the residual plot from the fit of the generalized additive model to the actual data and the others are the residual plot from the fit of the model to the null data. Which plot looks the most different?

Figure 5.

Both Manhattan plots show $-\mathrm{l}\mathrm{o}{\mathrm{g}}_{10}(P\text{-}value)$ of the genome-wide association analysis of withers height of sheep. Plot A is an informative plot while Plot B has extra annotations that enrich the narrative about the genome associations with the trait of interest. The horizontal dashed lines indicate where $-\mathrm{l}\mathrm{o}{\mathrm{g}}_{10}(P\text{-}value)$ is 5 and 7.

Figure 6.

Scheme of interactive graphics. The user interacts with the graphic by hovering or clicking. The graphic returns feedback to the user.

Figure 7.

Interactive visualization enabled by the plotly R package. Growth trajectories of 10 chickens over 21 d are shown. A: Displaying chick 7 data at the age of 16 d on hover. B: Comparing weight of 10 chicks at the age of 16 d. C: Single click on the chick 7 legend hides its trace. D: Double-click on the chick 7 legend isolates its trace. The interactive version of this plot is available at http://emitanaka.org/supp/anisci-datavis/#chick-plotly.

Figure 8.

Linking an interactive scatter plot and a bar chart to explore the Jersey data. The scatter plot on the left panel displays the relationship between fat yield deviation (x-axis) and protein yield deviation (y-axis). The bar chart on the right panel shows the frequency of cows classified according to their lactation numbers. A: Hovering on lactation 1 bar highlights lactation 1 cows in the scatter plot. B: Hovering on lactation 5 bar highlights lactation 5 cows in the scatter plot. The interactive version of the multiple linked plots is available at http://emitanaka.org/supp/anisci-datavis/#jersey-plotly.

Figure 9.

Scheme of Shiny applications. The user interacts or send events to the graphic or control widgets by mouse hovering or clicking. The user interface or graphic process user inputs and pass the code instructions to the server. The server modifies and update the data. The graphic returns feedback to the user.

Figure 10.

Shiny application for performing brushing in ggplot2 using the Dorper $\times$ Red Maasi lamb data. The scatter plot, box plot, and table are linked. (A) The user creates the blue-shaded rectangular box interactively in the scatter plot selecting the data points of interest. The box plot of weaning weight between sex and the table dynamically display information about those selected data points. (B) Brushing or dragging the rectangular box to another group of data points dynamically updates the box plot and the table. The interactive version of the Shiny application is available at https://chikudaisei.shinyapps.io/ggplot2-brushing/.

Figure 11.

Visualization of a dairy cow (A) and a group of pigs (B) using top-view depth images. Continuously varying heights are indicated by different colors.