Error in K-function Estimation

[Janjua1991]

Hi Jeremy. Thank you for updating and improving the spNetwork package.

I am facing some issues while running the k-function code. I am using projected coordinates for both line and point shapefiles. The error I am getting is :

Error in coordsLines[[nearest_line_index[x]]] : attempt to select less than one element in get1index

Can you please help me to resolve this error? Thanks

[JeremyGelb]

Hello There!

First of all, thank you for your interest in spNetwork! Could you send me a minimal sample of the data your are using and some code to replicate the bug?

[Janjua1991]

I have reduced the network size and when i run the code again, this time i got different error.

I was running following Code:

library(spNetwork) library(tmap) library(rgdal) library(sf)

Lines<- read_sf('D:/Data for jeremy/Sample Data/Sample_lines.shp') Points <- read_sf('D:/Data for jeremy/Sample Data/Sample_Points.shp')

tm_shape(Lines) + tm_lines("black") + tm_shape(Points) + tm_dots(col = "red", size = 0.1)

k_Incidents <- kfunctions(Lines, Points, start = 1, end = 5000, step = 50, width = 1000, nsim = 10, resolution = 50, verbose = FALSE, conf_int = 0.05)

This is the error i got this time: Error in data.frame(obs_k = k_vals, lower_k = k_stats[1, ], upper_k = k_stats[2, : arguments imply differing number of rows: 101, 100

In another attempt, I tried to fix network topology issues. but again in the end I got an issue. this was the error message I received when I was using the whole network and full point data set. Error: cannot allocate vector of size 783.2 Mb

Sample Data.zip

[JeremyGelb]

Ok, I verified your example and indeed you found a bug! Thank you for sending me the data to reproduce it. It is very appreciated.

I will correct it as soon as possible. If you need your results quickly, a slight modification in your code would do the trick : replace the value of start by 0:

library(spNetwork)
library(tmap)
library(sf)

Lines<- st_read('C:/Users/Gelb/Desktop/TEMP/test_error/Sample.Data/Sample_lines.shp')
Points <- st_read('C:/Users/Gelb/Desktop/TEMP/test_error/Sample.Data/Sample_Points.shp')

tm_shape(Lines) +
  tm_lines("black") +
  tm_shape(Points) +
  tm_dots(col = "red", size = 0.1)

k_Incidents <- kfunctions(Lines, Points,
                          start = 0, end = 5000, step = 50,
                          width = 1000, nsim = 10, resolution = 50,
                          verbose = TRUE, conf_int = 0.05)

I will close this issue when the bug will be properly corrected.

[Janjua1991]

Thank you for your quick response and guidance. I am facing another error as well, please guide me on this too: Error in igraph::distances(graph, v = snapped_events$vertex_id, to = snapped_events$vertex_id) : At core/paths/dijkstra.c:148 : Duplicate vertices in `to', this is not allowed, Invalid value

[Janjua1991]

I am providing here a complete code with data to replicate the results. I am also sharing the outcome showing the error:

Code

library(spNetwork)
library(tmap)
library(sf)

Lines<- st_read('D:/Data for jeremy/Datafile/Lines.shp')
Points<- st_read('D:/Data for jeremy/Datafile/Points.shp')

tm_shape(Lines) +
  tm_lines("black") +
  tm_shape(Points) +
  tm_dots(col = "red", size = 0.1)

k_results <- kfunctions(Lines, Points,
                          start = 0, end = 500, step = 10,
                          width = 20, nsim = 50, resolution = 50,
                          verbose = TRUE, conf_int = 0.05)

Outcome with error in last line

> library(spNetwork)
> library(tmap)
> library(sf)
> 
> Lines<- st_read('D:/Data for jeremy/Datafile/Lines.shp')
Reading layer `Lines' from data source `D:\Data for jeremy\Datafile\Lines.shp' using driver `ESRI Shapefile'
Simple feature collection with 19155 features and 10 fields
Geometry type: LINESTRING
Dimension:     XY
Bounding box:  xmin: -2125768 ymin: 4605261 xmax: -2112230 ymax: 4618447
Projected CRS: WGS 84 / UTM zone 15N
> Points<- st_read('D:/Data for jeremy/Datafile/Points.shp')
Reading layer `Points' from data source `D:\Data for jeremy\Datafile\Points.shp' using driver `ESRI Shapefile'
Simple feature collection with 5688 features and 2 fields
Geometry type: POINT
Dimension:     XY
Bounding box:  xmin: -2125460 ymin: 4605663 xmax: -2112540 ymax: 4618229
Projected CRS: WGS 84 / UTM zone 15N
> 
> tm_shape(Lines) +
+   tm_lines("black") +
+   tm_shape(Points) +
+   tm_dots(col = "red", size = 0.1)
> 
> 
> k_results <- kfunctions(Lines, Points,
+                           start = 0, end = 500, step = 10,
+                           width = 20, nsim = 50, resolution = 50,
+                           verbose = TRUE, conf_int = 0.05)
[1] "Preparing data ..."
[1] "Snapping points on lines ..."
[1] "Building graph ..."
Error in igraph::distances(graph, v = snapped_events$vertex_id, to = snapped_events$vertex_id) : 
  At core/paths/dijkstra.c:148 : Duplicate vertices in `to', this is not allowed, Invalid value
>

[JeremyGelb]

Thank you again for the detailed problem description! It is really appreciated.

The error is caused by points that share the same location when they are snapped on the network. I will provide an error message about this case in the next version. Currently, you could just set the parameter agg to a small value like 2 (meters) so that points in a two meters radius will be merged (and their weights added). This should be a reasonable approximation and the function will work.

Again, thank you very much for your participation in spNetwork's bugs resolving!

[JeremyGelb]

This has been corrected in the new dev version. I am closing the issue. Feel free to open a new one if you find another bug