CODEWITHSUNDEEP
rheatmap
Nazer
Nazer

Reputation: 3774

Fill in a gradient across longitude and latitude

I have a csv file with latitude, longitude, and radiation. The data looks like a grid and there about 100 points across the state. I know there are a lot of R mapping options out there, but it seems easiest to take what I have and try to make a gradient from point to point, so it looks like a heat map. All of the heat map options I see work off of density and don't look accurate when I use them with these data. Is there a function I can use to smooth out the colors between points?

Here is a reproducible example with a much smaller range of numbers:

lat    lon    value
40.5   -91.5  4513619
41.0   -95.5  4490658
41.5   -96.0  4453810
41.5   -94.5  4475373
42.0   -95.5  4432107  
42.0   -94.5  4416015   

ggplot(df, aes(x = lon, y = lat, color = value)) + 
  geom_point(size=8) + 
  coord_equal() + 
  xlab('Longitude') + 
  ylab('Latitude')

Upvotes: 2

Views: 454

Answers (1)

www
www

Reputation: 39174

We can use geom_tile from the ggplot2 package to create a heatmep.

library(ggplot2)

ggplot(df, aes(x = lon, y = lat, fill = value)) + 
  geom_tile(color = "black") +
  scale_fill_viridis_c() +
  coord_equal() + 
  xlab('Longitude') + 
  ylab('Latitude') +
  theme_bw() + 
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.border = element_blank())

enter image description here

Since you only provided six data points, the plot looks strange. Below I completed and expand your example dataset, and then use the same code to plot the data.

library(dplyr)
library(tidyr)

set.seed(123)

df2 <- df %>%
  complete(lat, lon = seq(-96, -91.5, by = 0.5)) %>%
  mutate(value = ifelse(is.na(value), sample(df$value, n(), replace = TRUE), value))

ggplot(df2, aes(x = lon, y = lat, fill = value)) + 
  geom_tile(color = "black") +
  scale_fill_viridis_c() +
  coord_equal() + 
  xlab('Longitude') + 
  ylab('Latitude') +
  theme_bw() + 
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.border = element_blank())

enter image description here

Finally, since you are working with a spatial dataset, let's just convert the data to raster and plot it using the mapview package.

library(sp)
library(sf)
library(raster)
library(mapview)

df_sp <- df2 %>%
  st_as_sf(coords = c("lon", "lat"), crs = 4326) %>%
  as("Spatial") 

r <- raster(ncol = 10, nrow = 4)
extent(r) <- extent(df_sp)
r <- rasterize(df_sp, r, df_sp$value)

mapview(r)

enter image description here

DATA

df <- read.table(text = "lat    lon    value
40.5   -91.5  4513619
41.0   -95.5  4490658
41.5   -96.0  4453810
41.5   -94.5  4475373
42.0   -95.5  4432107  
42.0   -94.5  4416015",
                 header = TRUE)

Upvotes: 1

Related Questions