CODEWITHSUNDEEP
rggplot2cartography
JulietteTC
JulietteTC

Reputation: 11

How to aggregate and plot large datasets on a grid

I'm working with a dataset that contains coordinates and fishing effort (HF).

Here is a sample of my dataset. The real dataset has 2985412 rows.

Year  Month Type             SI_LONG  SI_LATI  HF
2008  04    OTB_DEF_>=70_0   -4.88    46.84    0.617
2008  04    OTB_DEF_>=70_0   -4.91    46.83    0.4  
2008  04    OTB_DEF_>=70_0   -4.96    46.85    1    
2008  04    OTB_DEF_>=70_0   -5.01    46.84    1.02 
2008  04    OTB_DEF_>=70_0   -5.01    46.92    1    
2008  04    OTB_DEF_>=70_0   -4.95    46.98    1    
2008  04    OTB_DEF_>=70_0   -4.92    46.96    1.02 
2008  04    OTB_DEF_>=70_0   -4.97    46.93    1.02 
2008  04    OTB_DEF_>=70_0   -4.96    46.97    0.167
2008  04    OTB_DEF_>=70_0   -5       46.94    0.833

As there are a lot of points, I would like to aggregate my data on a grid with a resolution of 0.05 so I can map my data.

The end result should be something that looks like this but with way less points and a pixel regrouping the mean of several points that are adjacent.

Expected visualization

Expected vizualisation

Here is what I tried for now but it does not give me quite what I want. HF_filter is the name of the dataframe that I presented.

# Map and grid limits
grid_xmin <- -6
grid_xmax <- 0
grid_ymin <- 42
grid_ymax <- 48
grid_limit <- extent(c(grid_xmin,grid_xmax,grid_ymin,grid_ymax))

# Resolution of grid
grid <- raster(grid_limit)
res(grid) <- 0.05
crs(grid) <- "+proj=longlat +datum=WGS84"
# Create grid polygon
gridpolygon <- rasterToPolygons(grid)
gridpolygon$layer <- c(1:length(gridpolygon$layer))
gridpolygon_sf <- st_as_sf(gridpolygon)

#Transform to sf object
points_sf <- st_as_sf(HF_filter[,c("SI_LONG","SI_LATI","HF","LE_MET_level6","Year","Month")],
                      coords = c("SI_LONG","SI_LATI"),crs="+proj=longlat +datum=WGS84")
points_sf <- points_sf[st_intersects(points_sf,gridpolygon_sf) %>% lengths > 0,]
points_df <- st_join(points_sf,gridpolygon_sf) %>% 
  as.data.frame %>%
  dplyr::select(-geometry)

## Aggregate at the cell level
polyg_sf <- inner_join(gridpolygon_sf,points_df)%>% 
  group_by(layer) %>% 
  dplyr::summarise(fishing_effort = mean(HF))

#Plot
ggplot() +
  geom_sf(data=polyg_sf,aes(fill = fishing_effort,col=fishing_effort))+
  #geom_sf(data=points_sf,size=0.1)+
  geom_sf(data=mapBase)+
  coord_sf(xlim = c(-6,0), ylim = c(43,48), expand = FALSE)+
  scale_fill_distiller(palette = "Spectral")+
  scale_color_distiller(palette = "Spectral")

It gives me something close to what I want except, all pixels are not the same size or some are on top of each other. I would like something a little more "organized".

Current result

Current result

What would you suggest?

Upvotes: 1

Views: 53

Answers (0)

Related Questions