How to remove random diagonal lines from star map vizualizations produced by geom_sf?

Question

Context

I am in the process of developing a custom star map package called starBliss . While I thought I sucessfully figured out what I needed to from my question , it looks like there is some edge cases where things start to break.

For more background check out the GitHub Issue here

The Question

I am presently working on a function that takes a constellation lines shapefile and re-projects it based on longitude and latitude values with a Lambert Azimuthal Equal Area projection to create the the night sky of a given location at any given time.

A successful implementation can be seen in with the present starBliss package:

#devtools::install_github("benyamindsmith/starBliss")
library(ggplot2)
library(starBliss)
p<- plot_starmap(location= "Toronto, ON, Canada",
             date="2022-01-17",
             style="black",
             line1_text="Toronto",
             line2_text ="January 17th, 2023",
             line3_text="43.6532° N, 79.3832° W")
ggsave('toronto_black.png', plot = p, width = unit(10, 'in'), 
       height = unit(15, 'in'))

However this approach does run into problems. For example:

(image is cropped)

plot_starmap(
  location= "Caracas, Venezuela",
  date = as.Date("1991-03-17"),
  style = "black")

The above code creates some off diagonal lines (circled in red)

A reproducible example

I put together a function that will get the constellation lines data and transform it as it presently is. When I use that and plot in ggplot2 it with geom_sf the problem still exists.

library(tidyverse)
library(sf)
library(tidygeocoder)
library(lubridate)  

custom_starmap <- function(location,
                           date){
  
  # Formatting Date properly
  date<- as.Date(date)
  # Formatted date
  dt<- lubridate::ymd(date)
  # Get Latitude and Longitude for ProjString
  
  # For Latitude
  
  suppressMessages(
    capture.output(
      lat <- tibble(singlelineaddress = location) %>%
        geocode(address=singlelineaddress,method = 'arcgis') %>% .[["lat"]]
    )
  )
  
  # Reference date used for calculating longitude
  ref_date <- paste0(year(dt),"01","01",sep="-") %>% ymd()
  # Resulting longitude
  lon <- (-as.numeric(difftime(ref_date,dt, units="days"))/365)*360
  
  
  # The CRS
  projString <- paste0("+proj=laea +x_0=0 +y_0=0 +lon_0=",round(lon,4), " +lat_0=", round(lat,4))
  
  
  # Data Transformation
  flip <- matrix(c(-1, 0, 0, 1), 2, 2)
  
  hemisphere <- st_sfc(st_point(c(lon, lat)), crs = 4326) %>%
    st_buffer(dist = 1e7) %>%
    st_transform(crs = projString)
  
  # Data source for constellation lines
  
  url1 <- "https://raw.githubusercontent.com/benyamindsmith/starBliss/main/data/constellations.lines.json"
  
  # Reading Data
  invisible(
    capture.output(
      constellation_lines_sf <- invisible(st_read(url1, stringsAsFactors = FALSE)) %>%
        st_wrap_dateline(options = c("WRAPDATELINE=YES", "DATELINEOFFSET=360")) %>%
        st_transform(crs = projString) %>%
        st_intersection(hemisphere) %>%
        filter(!is.na(st_is_valid(.))) %>%
        mutate(geometry = geometry * flip)
    )
  )
  
  
  st_crs(constellation_lines_sf) <- projString
  
  return(constellation_lines_sf)
}

# The data

df<-custom_starmap(location= "Caracas, Venezuela",
                   date = as.Date("1991-03-17"))

df

> Simple feature collection with 49 features and 2 fields
Geometry type: GEOMETRY
Dimension:     XY
Bounding box:  xmin: 8967611 ymin: -8898251 xmax: -8714977 ymax: 9004400
CRS:           +proj=laea +x_0=0 +y_0=0 +lon_0=73.9726 +lat_0=10.488
First 10 features:
    id rank                       geometry
1  And    1 MULTILINESTRING ((3542468 3...
2  Ant    3 LINESTRING (-6234955 -52010...
3  Aqr    2 MULTILINESTRING ((8967611 -...
4  Ari    1 LINESTRING (3098546 2071855...
5  Aur    1 MULTILINESTRING ((-1307725 ...
6  Cae    3 LINESTRING (557848.4 -59059...
7  Cam    2 MULTILINESTRING ((-24783.5 ...
8  Cnc    2 MULTILINESTRING ((-6264812 ...
9  CMa    1 MULTILINESTRING ((-2356827 ...
10 CMi    2 LINESTRING (-4432439 -32157..

When I plot this data the lines in question can be seen:

df %>% ggplot()+geom_sf()

(circled in red for clarity)

How do I fix this? Is there an issue with the format of the CRS that I am using? Or do I need to crop the lines?

Answer 1

I think it is safer to use s2 for this type of exercise:

library(tidyverse)
library(sf)
#> Linking to GEOS 3.9.3, GDAL 3.5.2, PROJ 8.2.1; sf_use_s2() is TRUE
library(tidygeocoder)
library(lubridate)  
#> Loading required package: timechange
#> 
#> Attaching package: 'lubridate'
#> The following objects are masked from 'package:base':
#> 
#>     date, intersect, setdiff, union

custom_starmap <- function(location,
                           date){
  
  # Formatting Date properly
  date<- as.Date(date)
  # Formatted date
  dt<- lubridate::ymd(date)
  # Get Latitude and Longitude for ProjString
  
  # For Latitude
  
  suppressMessages(
    capture.output(
      lat <- tibble(singlelineaddress = location) %>%
        geocode(address=singlelineaddress,method = 'arcgis') %>% .[["lat"]]
    )
  )
  
  # Reference date used for calculating longitude
  ref_date <- paste0(year(dt),"01","01",sep="-") %>% ymd()
  # Resulting longitude
  lon <- (-as.numeric(difftime(ref_date,dt, units="days"))/365)*360
  
  
  # The CRS
  projString <- paste0("+proj=laea +x_0=0 +y_0=0 +lon_0=",round(lon,4), " +lat_0=", round(lat,4))
  
  
  # Data Transformation
  flip <- matrix(c(-1, 0, 0, 1), 2, 2)
  
  # Hemisphere with s2
  hemisphere <- s2::s2_buffer_cells(
    s2::as_s2_geography(paste0("POINT(", lon, " ", lat, ")")),
    distance = 1e7,
    max_cells = 5000)

  
  # Data source for constellation lines
  
  url1 <- "https://raw.githubusercontent.com/benyamindsmith/starBliss/main/data/constellations.lines.json"
  
  # Reading Data
  invisible(
    capture.output(
      constellation_lines_sf <- invisible(st_read(url1, stringsAsFactors = FALSE)) %>%
        st_wrap_dateline(options = c("WRAPDATELINE=YES", "DATELINEOFFSET=360")) %>%
        # Use s2 for the cut
        st_as_s2() %>%
        s2::s2_intersection(hemisphere) %>%
        # Back to sf
        st_as_sf() %>%
        st_transform(crs = projString) %>%
        filter(!is.na(st_is_valid(.))) %>%
        mutate(geometry = geometry * flip) %>%
        # Filter if empty, since the cut can produce empty geometries
        filter(!st_is_empty(.))

    )
  )
  
  
  st_crs(constellation_lines_sf) <- projString
  
  return(constellation_lines_sf)
}

# The data

df<-custom_starmap(location= "Caracas, Venezuela",
                   date = as.Date("1991-03-17"))

df
#> Simple feature collection with 48 features and 0 fields
#> Geometry type: GEOMETRY
#> Dimension:     XY
#> Bounding box:  xmin: -8700015 ymin: -8913303 xmax: 8922028 ymax: 8998639
#> CRS:           +proj=laea +x_0=0 +y_0=0 +lon_0=73.9726 +lat_0=10.488
#> First 10 features:
#>                          geometry
#> 1  MULTILINESTRING ((3542468 3...
#> 2  LINESTRING (-6234955 -52010...
#> 3  MULTILINESTRING ((8922028 -...
#> 4  LINESTRING (3098546 2071855...
#> 5  MULTILINESTRING ((-1307725 ...
#> 6  LINESTRING (557848.4 -59059...
#> 7  MULTILINESTRING ((-24783.5 ...
#> 8  MULTILINESTRING ((-6264812 ...
#> 9  MULTILINESTRING ((-2356827 ...
#> 10 LINESTRING (-4432439 -32157...

ggplot(df) +
  geom_sf()

^{Created on 2023-01-23 with reprex v2.0.2}