Figure 7.6 Space-time paths for 100 Beijing taxis

figures

code

Taking advantage of the web format, here I show an interactive version of this figure.

Code

library(dplyr)
library(timeDate)
library(gg3D)
library(plotly)

We filter the data so they are all within 0.25 decimal degrees of the mean location of all points (I know, I know… decimal degrees are terrible, but the point here is the time dimension, not the geographical coordinates).

Code

taxis <- read.table('taxis100.txt', sep=',') |>
  rename(id = V1, t = V2, lon = V3, lat = V4)

lon.mean <- mean(taxis$lon)
lat.mean <- mean(taxis$lat)

taxis.bj <- taxis |>
  mutate(time = as.double(timeDate(t))) |>
  filter(abs(lon - lon.mean) < .25,
         abs(lat - lat.mean) < .25)

plotly makes a nice interactive plot, without too much fuss.

Code

plot_ly(group_by(taxis.bj, id), 
        x = ~lat, y = ~lon, z = ~t, color = ~id, 
        type = 'scatter3d', mode = 'lines', asp = 1, lwd = 0.5)