This function plots a previously calculated trend.
plot_trend(df, column_name, maptype = "stamen_toner_lite", showmap = TRUE)Data frame containing at least 4 column:
lat (latitude), lon (longitude), slope and an additional user-defined column
column_name.
name of the column to use for grouping the results.
set to FALSE to avoid plotting the map when running the function
maptype, was need to choose the stamenmap type, now useless since stamenmap are no longer reachable
Two plots, the first showing the distribution of the
trend over a map, based on the slope of the linear model that describes the
trend. The second plot shows a boxplot of the slope grouped based on the
column column_name and slope can be user-defined
(notice that in the plot the very extreme slope values are not
displayed to avoid skewed visualisations).
The function relies on the `ggmap` package for the map, and this package has in time gone through many changes due to changes in API of map providers. Currently to be able to create the map one needs to register to the stadiamaps serice. More information at ?ggmap::register_stadiamaps().
if (FALSE) {
# some fake data around London
df <- data.frame(lat = 51.5+runif(40,-0.3,0.3),
lon = 0+runif(40, -0.3,0.3),
slope = rnorm(40, c(rep(-0.4,20),rep(0.4,20))),
g = factor(c(rep("a",20), rep("b",20))))
theplots <- plot_trend(df, "g", maptype = "terrain-background")
theplots$A # map
theplots$B + labs(subtitle = "Use ggplot usual commands to modify the plots") # boxplots
}