r - 使用 mapdeck 的 add_trips 函数创建数据层时出错

Question

尝试为 mapdeck 使用新的 add_trips 层。下面的示例代码和错误。感谢任何帮助。谢谢。

我正在使用的数据：

> class(msi)
[1] "sf"         "tbl_df"     "tbl"        "data.frame"

> head(msi$geometry)
Geometry set for 6 features 
geometry type:  POINT
dimension:      XYZ
bbox:           xmin: 2.94486 ymin: 51.34172 xmax: 3.21298 ymax: 51.42742
epsg (SRID):    4326
proj4string:    +proj=longlat +datum=WGS84 +no_defs
First 5 geometries:
POINT Z (2.94486 51.42742 1573824004)
POINT Z (3.2062 51.35052 1573827317)
POINT Z (3.21298 51.34172 1573830334)
POINT Z (3.21298 51.34175 1573834830)
POINT Z (3.21297 51.34173 1573838433)

地图甲板代码：

key = "mykeyhere"

plot <- mapdeck( token = key, style = 'mapbox://styles/mapbox/dark-v9',
                         pitch = 30,
                         width="100%",
                         zoom=15) %>%
  add_trips(
    data = msi
  )

我得到的错误：

Error in rcpp_path_geojson(data, l, geometry_column, digits, "trips") : Error creating data layer

Answer 1

Notes on your example

1. The trips layer requires LINESTRINGs, not POINTs
2. It also needs both Z (elevation) and M (time) attributes (the lines are rendered in space and time)

Other details

I wrote the trips layer with this pull-request into the sf library in mind, so I've designed it to automatically read the z_range and m_range of the sfc object.

I also have the sfheaders library which can construct sf objects which include these attributes.

I am using the github versions of sf and mapdeck

remotes::install_github("r-spatial/sf")
remotes::install_github("SymbolixAU/mapdeck")

Working Example

You haven't included a reproducible data set in your question, so I'm going to use my own.

In this example I

1. Take a sf LINESTRING object
2. Convert into a data.table of coordinates
3. Append Z and M columns
4. Convert back to sf LINESTRING object
5. Plot trips

library(mapdeck)
library(sf)
library(sfheaders)
library(data.table) 

## 'roads' is an sf object included in mapdeck
sf_roads <- mapdeck::roads

## you don't need this, I only use it later in the plot so I can subset only 'long' 
## roads, so the gif is below 2mb!
sf_roads$length <- as.numeric( sf::st_length( sf_roads ) )
sf_roads$id <- 1:nrow( sf_roads )  ## for joining the 'length' back on

## grab the road coordinates
dt <- sf::st_coordinates( sf_roads )
dt <- as.data.table( dt )

## st_coordinates returns a L1 'id', which is equivalent to our 'id' above

## Need Z and M attributes
dt[, Z := 0 ]   ## can be actual elevation if you have it
dt[, M := 1:.N, by = .(L1)]   ## this 'M' is your timestamp

## convert back to sf LINESTRING
sf_trips <- sfheaders::sf_linestring(
  obj = dt
  , x = "X"
  , y = "Y"
  , z = "Z"
  , m = "M"
  , linestring_id = "L1"
)

## You don't need this bit, where I join the 'length' back on.
sf_roads$geometry <- NULL

sf_trips <- merge(
  x = sf_trips
  , y = sf_roads
  , by.x = "L1"
  , by.y = "id"
)



mapdeck(
  style = mapdeck_style("dark")
) %>%
  add_trips(
    data = sf_trips[ sf_trips$length > 500, ]  ## only showing 'long' roads
    , stroke_colour = "L1"
    , stroke_width = 100
    , opacity = 0.8
    , animation_speed = 10
    , trail_length = 50
    , palette = 
  )