CODEWITHSUNDEEP
rdatetimeggplot2ggmap
Rory Shaw
Rory Shaw

Reputation: 851

Using date time data with ggplot scale_colour_gradient

I am plotting some time series GPS coordinates using ggmap and ggplot. I want to visualise the time series by creating a colour gradient. I have had a couple of attempts so far as shown below.

My data can be accessed here

Import data

Dec7 = read.csv("7-12-15.csv", header = TRUE, stringsAsFactors = FALSE)
Dec7$timestamp <- as.Date(Dec7$timestamp)

head(Dec7)
                  X_id seq_id  timestamp       lon address      lat rssi sensor gps_quality batt_v
1 56656ecd0dd8e408d8c2e43f     71 2015-12-07 -3.780899     208 53.20252  -63    gps           1   3274
2 56656ed20dd8e408d8c2e440     72 2015-12-07 -3.780958     208 53.20246  -63    gps           1   3274
3 56656edc0dd8e408d8c2e441     73 2015-12-07 -3.780967     208 53.20246  -65    gps           1   3274
4 56656ee60dd8e408d8c2e442     74 2015-12-07 -3.780968     208 53.20242  -64    gps           1   3274
5 56656ef10dd8e408d8c2e443     75 2015-12-07 -3.780997     208 53.20240  -64    gps           1   3274
6 56656efa0dd8e408d8c2e446     76 2015-12-07 -3.780965     208 53.20243  -64    gps           1   3274


 str(Dec7)
data.frame':    22420 obs. of  10 variables:
 $ X_id       : chr  "56656ecd0dd8e408d8c2e43f" "56656ed20dd8e408d8c2e440" "56656edc0dd8e408d8c2e441" "56656ee60dd8e408d8c2e442" ...
 $ seq_id     : int  71 72 73 74 75 76 77 78 86 87 ...
 $ timestamp  : Date, format: "2015-12-07" "2015-12-07" "2015-12-07" "2015-12-07" ...
 $ lon        : num  -3.78 -3.78 -3.78 -3.78 -3.78 ...
 $ address    : num  208 208 208 208 208 208 208 208 208 208 ...
 $ lat        : num  53.2 53.2 53.2 53.2 53.2 ...
 $ rssi       : int  -63 -63 -65 -64 -64 -64 -64 -63 -64 -64 ...
 $ sensor     : chr  "gps" "gps" "gps" "gps" ...
 $ gps_quality: int  1 1 1 1 1 1 1 1 1 1 ...
 $ batt_v     : int  3274 3274 3274 3274 3274 3274 3274 3274 3274 3274 ...

I have classed timestamp as.Date as I am aware that this can then be passed successfully into scale_colour_gradient within the ggplot call as follows:

mapImageData <- get_googlemap(center = c(lon = median(Dec7$lon), 
                                     lat = median(Dec7$lat)), zoom = 17,
                          size = c(500, 500),
                          maptype = c("satellite"))   

sheep_hiraetlyn_Dec7_map = ggmap(mapImageData,extent = "device") + 
  geom_point(aes(x = lon,y = lat, color=timestamp),
         data = Dec7, size = 1, pch = 20) + 
  scale_color_gradient(trans = "date", low="red", high="blue")

This produces the following map:

enter image description here

As you can see the colour gradient is discrete rather than the desired continuous gradient - presumably this is because it categorizes the timestamp into days? Also The legend labels consist of 2 overlayed labels so are not clear. I have tried using as.POSIXct but this cannot be passed to trans. I have also used as.Integer, which creates a nice colour gradient but the legend cannot be interpreted in terms of date/time.

Any ideas how I can get round this problem?

Thanks

Upvotes: 3

Views: 2975

Answers (1)

Thierry
Thierry

Reputation: 18487

Convert the POSIXct timestamps to integer and define the breaks and labels manually

Dec7$time <- as.integer(Dec7$timestamp)
labels <- pretty(Dec7$timestamp, 5)
ggmap(mapImageData,extent = "device") + 
  geom_point(
    aes(x = lon,y = lat, color=time), 
    data = Dec7, size = 1, pch = 20
  ) + 
  scale_color_gradient(
    low="red", high="blue", 
    breaks = as.integer(labels), 
    labels = format(labels, "%m/%d %H:%M")
  )

Upvotes: 3

Related Questions