Time series aim to study the evolution of one
or several variables through time. This section gives examples using R
. A focus is
made on the tidyverse
: the lubridate
package is indeed your best friend to
deal with the date format, and ggplot2
allows to plot it efficiently. The dygraphs
package
is also considered to build stunning interactive charts. Note that this
online course is dedicated to timeseries analysis with R.
Date
format? You will ♥ the lubridate
package.Building time series requires the time variable to be at the date
format. The first step of your analysis must be to double check that R read your data correctly, i.e. at the date
format. This is possible thanks to the str()
function:
Getting this date
format can be a pain, and the lubridate
package is such a life saver. It offers several function which name are composed by 3 letters: year (y
), month (m
) and day (d
). Example:
Note: this course teaches how to deal with dates and times more in depth.
ggplot2
ggplot2
offers great features when it comes to visualize time series. The date
format will be recognized automatically, resulting in neat X axis labels. The scale_x_data()
makes it a breeze to customize those labels. Last but not least, plotly
can turn the resulting chart interactive in one more line of code.
plotly
The ggplotly()
function of the plotly
library makes it a breeze to build an interactive version. Try to hover circles to get a tooltip, or select an area of interest for zooming. Double click to reinitialize.
dygraph
The dygraphs package is a html widget. It allows to make interactive time series chart: you can zoom and hover data points to get additional information. Start by reading the chart #316 for quick introduction and input description. Then, the graph #317 gives an overview of the different types of charts that are offered. To go further, check the graph #318 (interactive version below).
The dygraph
package offers zooming, hovering, minimaps and much more. Try it on the example below!
Heatmaps can be a very good alternative to visualize time series, especially when the time frame you study is repeating, like weeks. Here is a customized example, but visit the heatmap section for more.
Code Heatmap sectionlatticeExtra
or ggplot2
Warning: a dual Y axis line chart represents the evolution of 2 series, each plotted according to its own Y scale. This kind of chart must be avoided, since playing with Y axis limits can lead to completely different conclusions. Visit data-to-viz for more info.
Why you should avoid itThe web is full of astonishing R charts made by awesome bloggers. The R graph gallery tries to display some of the best creations and explain how their source code works. If you want to display your work here, please drop me a word or even better, submit a Pull Request!