Plotly quickstart guide

This document shows how plotly can help statistics and data scientists interactively select or deselect data, filter or zoom it, primarily when they work with large or dense data.
Plotly is a strength visualization package because it creates highly interactive and visually appealing charts, bars and graphs.
In this vignette, we will start with an introduction of plotly then we will examine the features of plotly charts as bellow;

Introduction to Plotly

According to this website, plotly is an open-source R package for creating interactive web-based graphs via the open-source JavaScript graphing library plotly.js. Plotly is free and can be used offline, without being connected to the internet [2].

Install and Download Plotly

There are two ways for downloading plotly:

Plotly package can be easily installed in CRAN by using install.package() function and loaded from library().

install.packages("plotly")  
library("plotly")

Plotly can install from GitHub through the devtools R package.

devtools::install_github("ropensci/plotly")

Download Dataset

To create basic graphs, we use the Salaries dataset from this website.

#> # A tibble: 397 x 6
#>    rank      discipline yrs.since.phd yrs.service sex    salary
#>    <chr>     <chr>              <dbl>       <dbl> <chr>   <dbl>
#>  1 Prof      B                     19          18 Male   139750
#>  2 Prof      B                     20          16 Male   173200
#>  3 AsstProf  B                      4           3 Male    79750
#>  4 Prof      B                     45          39 Male   115000
#>  5 Prof      B                     40          41 Male   141500
#>  6 AssocProf B                      6           6 Male    97000
#>  7 Prof      B                     30          23 Male   175000
#>  8 Prof      B                     45          45 Male   147765
#>  9 Prof      B                     21          20 Male   119250
#> 10 Prof      B                     18          18 Female 129000
#> # ... with 387 more rows

Create Plotly

In this vignette, we want to show two different approaches for creating Plotly:

Use plotly directly by using plot_ly()
Converting ggplot2 to plotly by using ggplotly() function

plot_ly()

Plot_ly() is a function of the plotly package which delivers a direct interface to plotly [1]. It has various arguments such as color, type, size, symbol and e.g.

Plot_ly(data, x, y, type, mode, color, size,…)

data A data frame
X, Y Axis values
type Specifies the type of plot such as scatter, bar, histogram, heatmap, box plots e.g.
Color The colour of data point, use the function of I() to avoid scaling
Size Specifies the size of data points, use via I() to avoid altogether (e.g., size = I(30)).

Organize Plotly attributes

There are two categories for describing plotly’s attributes:
- traces (describe a single series of data in a graph)
- layout (apply to the rest of the graph, e.g. title, xaxis, yaxis)

The following example shows how trace_add() and layout() are inlined with links to each attributes reference section.

Plot_ly(data, x, y, type, mode, color, size,….) %>%
add_trace(x, y, mode, list, color) %>%
layout(title, xaxis, yaxis)

Create Scatter Plot

Figure 1 shows the basic scatter plot by using the plot_ly function.

pal <- c("darkolivegreen3", "dodgerblue3", "deeppink3")
fig <- plot_ly(data = Salaries, x= ~yrs.service, y= ~salary, color = ~rank, colors = pal)
fig

Figure 1: Interactive scatter plot by plot_ly

By clicking on legend entries, we can simply download the plot as a png, hide or show traces, click-and-drag on the chart to zoom in or zoom out, double-click to autoscale, shift-and-drag to the pan. Also, we can select and deselect university ranks on the graph.

Create Histogram

Figure 2 shows the basic interactive histogram. As codes show, we can have a histogram graph by adding type = “histogram” argument. If we want to have other graph types, we need to change the type attribute input.

pal = c("deeppink3","dodgerblue3")
axx <- list( title = "Years")
axy <- list(title = "Amount")
fig <- plot_ly(data = Salaries, x= ~yrs.service, type = "histogram")
fig <- fig %>% add_trace( type="histogram", x=~yrs.since.phd)
fig

Figure 2: Interactive histogram by plot_ly

The histogram of “yrs.since.phd” added to the existing histogram by inputting add_trace().

Create 3D Plot_ly

Figure 3 shows the 3D interactive graph. In this example, we set axes names by linking layout().

axx <- list( title = "Years.service")
axy <- list(title = "Salary")
axz <- list(title = "Rank")

fig <- plot_ly(data = Salaries, x= ~yrs.service, y= ~salary, z = ~rank, type = 'mesh3d', aspectmode='data')
fig <- fig %>% layout(scene = list(xaxis=axx,yaxis=axy,zaxis=axz))
fig

Figure 3: Interactive 3D Plot_ly

ggplotly()

ggplotly() is a function which is converted ggplot2 graphs to interactive plotly objects. It is a quick way for adding interactivity to existing ggplot2 charts [1].

ggplotly(p, width = NULL, height = NULL, dynamicTicks = FALSE, layerData = 1, originalData = TRUE, source = "A", ...)

p ggplot object
width Width of the plot (optional).
height Height of the plot (optional).
Argument p is mandatory, but other arguments are optional. For more information about other arguments, please visit this website.


P <-ggplot(data = Salaries, aes(x= yrs.service, y= salary))+  geom_point(aes(color = rank)) + geom_smooth(method = "lm", se = FALSE)+ labs(x="years of service",y="Salary") + theme_classic()
ggplotly(P)