[英]Using R to identify nearest point to a location and calculate the distance between them along a network/road
I have a series of locations (Points_B) and would like to find the closest point to them from a different set of points (Points_A) and the distance between them in kms. 我有一系列位置(Points_B),想从一组不同的点(Points_A)中找到最接近它们的点,以及它们之间的距离(以kms为单位)。 I can do this as the crow flies but cannot work out how to do the same along a road network (the 'Roads' object in the code). 我可以在乌鸦飞翔时做到这一点,但无法弄清楚如何在道路网(代码中的“ Roads”对象)上做同样的事情。 The code I have so far is a follows: 到目前为止,我的代码如下:
library(sp)
library(rgdal)
library(rgeos)
download.file("https://dl.dropboxusercontent.com/u/27869346/Road_Shp.zip", "Road_Shp.zip")
#2.9mb
unzip("Road_Shp.zip")
Roads <- readOGR(".", "Subset_Roads_WGS")
Points_A <- data.frame(ID = c("A","B","C","D","E","F","G","H","I","J","K","L"), ID_Lat = c(50.91487, 50.92848, 50.94560, 50.94069, 50.92275, 50.94109, 50.92288, 50.92994, 50.92076, 50.90496, 50.89203, 50.88757), ID_Lon = c(-1.405821, -1.423619, -1.383509, -1.396910, -1.441801, -1.459088, -1.466626, -1.369458, -1.340104, -1.360153, -1.344662, -1.355842))
rownames(Points_A) <- Points_A$ID
Points_B <- data.frame(Code = 1:30, Code_Lat = c(50.92658, 50.92373, 50.93785, 50.92274, 50.91056, 50.88747, 50.90940, 50.91328, 50.91887, 50.92129, 50.91326, 50.91961, 50.91653, 50.90910, 50.91432, 50.93742, 50.91848, 50.93196, 50.94209, 50.92080, 50.92127, 50.92538, 50.88418, 50.91648, 50.91224, 50.92216, 50.90526, 50.91580, 50.91203, 50.91774), Code_Lon = c(-1.417311, -1.457155, -1.400106, -1.374250, -1.335896, -1.362710, -1.360263, -1.430976, -1.461693, -1.417107, -1.426709, -1.439435, -1.429997, -1.413220, -1.415046, -1.440672, -1.392502, -1.459934, -1.432446, -1.357745, -1.374369, -1.458929, -1.365000, -1.426285, -1.403963, -1.344068, -1.340864, -1.399607, -1.407266, -1.386722))
rownames(Points_B) <- Points_B$Code
Points_A_SP <- SpatialPoints(Points_A[,2:3])
Points_B_SP <- SpatialPoints(Points_B[,2:3])
Distances <- (gDistance(Points_A_SP, Points_B_SP, byid=TRUE))*100
Points_B$Nearest_Points_A_CF <- colnames(Distances)[apply(Distances,1,which.min)]
Points_B$Distance_Points_A_CF <- apply(Distances,1,min)
The output I am after would be two additional columns in 'Points_B' with 1) having the nearest Point A object ID along the road network and 2) having the distance along the network in km . 我得到的输出将是'Points_B'中的另外两列,其中1)沿道路网络具有最近的Point A对象ID , 2)沿网络的距离为km 。 Any help would be appreciated. 任何帮助,将不胜感激。 Thanks. 谢谢。
I've been working on this kind of problem all day. 我整天都在研究这种问题。 Try mapdist()
in the ggmap
package and see if this works: 在ggmap
包中尝试mapdist()
,看看是否ggmap
:
library(dplyr)
library(ggmap)
#Your data
Points_A <- data.frame(ID = c("A","B","C","D","E","F","G","H","I","J","K","L"), ID_Lat = c(50.91487, 50.92848, 50.94560, 50.94069, 50.92275, 50.94109, 50.92288, 50.92994, 50.92076, 50.90496, 50.89203, 50.88757), ID_Lon = c(-1.405821, -1.423619, -1.383509, -1.396910, -1.441801, -1.459088, -1.466626, -1.369458, -1.340104, -1.360153, -1.344662, -1.355842))
Points_B <- data.frame(Code = 1:30, Code_Lat = c(50.92658, 50.92373, 50.93785, 50.92274, 50.91056, 50.88747, 50.90940, 50.91328, 50.91887, 50.92129, 50.91326, 50.91961, 50.91653, 50.90910, 50.91432, 50.93742, 50.91848, 50.93196, 50.94209, 50.92080, 50.92127, 50.92538, 50.88418, 50.91648, 50.91224, 50.92216, 50.90526, 50.91580, 50.91203, 50.91774), Code_Lon = c(-1.417311, -1.457155, -1.400106, -1.374250, -1.335896, -1.362710, -1.360263, -1.430976, -1.461693, -1.417107, -1.426709, -1.439435, -1.429997, -1.413220, -1.415046, -1.440672, -1.392502, -1.459934, -1.432446, -1.357745, -1.374369, -1.458929, -1.365000, -1.426285, -1.403963, -1.344068, -1.340864, -1.399607, -1.407266, -1.386722))
#Combine coords into one field (mapdist was doing something funny with the commas so I had to specify "%2C" here)
Points_A$COORD <- paste(ID_Lat, ID_Lon, sep="%2C")
Points_B$COORD <- paste(Code_Lat, Code_Lon, sep="%2C")
#use expand grid to generate all combos
get_directions <- expand.grid(Start = Points_A$COORD,
End = Points_B$COORD,
stringsAsFactors = F,
KEEP.OUT.ATTRS = F) %>%
left_join(select(Points_A, COORD, ID), by = c("Start" = "COORD")) %>%
left_join(select(Points_B, COORD, Code), by = c("End" = "COORD"))
#make a base dataframe
route_df <- mapdist(from = get_directions$Start[1],
to = get_directions$End[1],
mode = "driving") %>%
mutate(Point_A = get_directions$ID[1],
Point_B = get_directions$Code[1])
#get the rest in a for-loop
start <- Sys.time()
for(i in 2:nrow(get_directions)){
get_route <- mapdist(from = get_directions$Start[i],
to = get_directions$End[i],
mode = "driving") %>%
mutate(Point_A = get_directions$ID[i],
Point_B = get_directions$Code[i])
route_df <<- rbind(route_df, get_route) #add to your original file
Sys.sleep(time = 1) #so google doesn't get mad at you for speed
end <- Sys.time()
print(paste(i, "of", nrow(get_directions),
round(i/nrow(get_directions),4)*100, "%", sep=" "))
print(end-start)
}
#save if you want
write.csv(route_df, "route_df.csv", row.names = F)
#Route Evaluation
closest_point <-route_df %>%
group_by(Point_A) %>%
filter(km == min(km)) %>%
ungroup()
I'm still kind of new at this so there may be a better way to do the data wrangling. 我对此还是有点陌生,因此可能会有更好的方法来处理数据。 Hope this helps & good luck 希望这有帮助,祝你好运
The packages igraph, osmr and walkalytics all seem to provide this functionality these days. 如今,igraph,osmr和walkalytics软件包似乎都提供了此功能。 Mode-specific routing networks exist (in various degrees of functionality). 存在特定于模式的路由网络(具有不同程度的功能)。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.