I'm doing a research about climate change and I want to extract the CMIP6 information using R but I don't get the results I want. This is the code.
library(raster)
library(ncdf4)
long_lat <- read.csv("C:/Users/USUARIO/Desktop/Tesis/Coordenadas - copia.csv", header = T)
raster_pp <- raster::brick("pr_day_ACCESS-CM2_historical_r1i1p1f1_gn_2000.nc")
sp::coordinates(long_lat) <- ~XX+YY
raster::projection(long_lat) <- raster::projection(raster_pp)
points_long_lat <- raster::extract(raster_pp[[1]], long_lat, cellnumbers = T)[,1]
data_long_lat <- t(raster_pp[points_long_lat])
The cordinates I've used are the following: -77.7754,-9.5352 (latitude, longitude), but I've tried these as well and still doesn't work 102.2246,-9.5352. Thanks for your response.
Related
I try to sew/merge/bind two LINESTRINGs together (spanish coast + french coast) to have the full map of the whole coast. I found all the shpaefiles here:
https://www.marineregions.org/gazetteer.php?p=details&id=3417
https://www.marineregions.org/gazetteer.php?p=details&id=19888
Import and plot the data
frenchCoast_CoteBanyuls <- st_read("coasts_subnational/coasts_subnational.shp") %>%
st_geometry() #
plot(frenchCoast_CoteBanyuls)
spainCoast_CoteBanyuls <- st_read("coasts_subnational SPAIN/coasts_subnational.shp")%>%
st_geometry() #
spainCoast_CoteBanyuls <- st_cast(x = spainCoast_CoteBanyuls, to = "LINESTRING")
plot(spainCoast_CoteBanyuls, add = T)
This map and features seem ok.
According to all the posts I found, I tried to merge the two shapefiles together:
CoteBanyuls_c <- c(frenchCoast_CoteBanyuls, spainCoast_CoteBanyuls)
# CoteBanyuls_rbind <- rbind(frenchCoast_CoteBanyuls, spainCoast_CoteBanyuls)
# CoteBanyuls <- st_union(CoteBanyuls)
# CoteBanyuls <- union(CoteBanyuls)
# CoteBanyuls <- bind(CoteBanyuls)
plot(CoteBanyuls)
the c() function gives a shapefile that does not have the good dimensions (see below) and none of these solutions does plot the good shape of the two coast merged (seems to plot only the spanish coast).
What did I do wrong ? I do not understand why these solutions are not working...
Do you have some ideas ?
Thanks in advance !
Charlotte
I think the issue comes from extracting only the geometry attributes before you combine them. Try this:
frenchCoast_CoteBanyuls <- st_read("coasts_subnational/coasts_subnational.shp")
spainCoast_CoteBanyuls <- st_read("coasts_subnational SPAIN/coasts_subnational.shp")
combined_coast <- rbind(spainCoast_CoteBanyuls, frenchCoast_CoteBanyuls)
mapview::mapview(combined_coast)
For scientific analysis, I am following this tutorial:
https://uw-madison-microbiome-hub.github.io/Microbiome_analysis_in-_R/
Specifically, I would like to use this script to make permanova:
library(vegan)
library(qiime2R)
library(phyloseq)
library(ggplot2)
# import data
ASVs <- read_qza("/path/table.qza")
metadata <- read.table("/path/metadata16S.tsv", , sep='\t', header=T, row.names=1, comment="")
metadata <- metadata[-1,]
tree <- read_qza("/path/rooted-tree.qza")
taxonomy <- read_qza("/path/taxonomy.qza")
tax_table <- do.call(rbind, strsplit(as.character(taxonomy$data$Taxon), ";"))
colnames(tax_table) <- c("Kingdom","Phylum","Class","Order","Family","Genus","Species")
rownames(tax_table) <- taxonomy$data$Feature.ID
# phyloseq object
physeq <- phyloseq(
otu_table(ASVs$data, taxa_are_rows = T),
phy_tree(tree$data),
tax_table(tax_table),
sample_data(metadata)
)
# convert ot relative abundance
physeq_rel <- microbiome::transform(physeq, "compositional")
physeq.ord.wuni <- ordinate(physeq_rel, "PCoA", "unifrac", weighted=T)
b.div.wuni <- plot_ordination(physeq_rel, physeq.ord.wuni, type= "samples", color= "MyParameter") + geom_point(size=3)
b.div.wuni <- b.div.wuni + stat_ellipse() + ggtitle("Weighted Unifrac") + theme_classic()
print(b.div.wuni)
The problem is that after running this script, I get purely random data every time I run it, but only on my Linux, while if I run it on other computers with a window OS, the results are correct!
What could be causing the problem?
The only clue I have is that during the installation of ggplot2 I had problems regarding libmkl_rt.so.1, already present in conda, but which R could not find. I solved the problem by renaming the files as suggested in https://github.com/igraph/rigraph/issues/275#issuecomment-947940126
I have no idea what to do ...
Thank you for your help
Andrea
I am trying to merge a shapefile and .csv file to make a map of election results. I can make a graph when I load the shapefile, but as soon as I merge it with the .csv, it says "Error in plot.window(...) : need finite 'xlim' values".
I have been reading online, and I think maybe I have to merge the csv file to the shapefile (I have been merging the shapefile to the csv file). However, the csv file (which contains the election results) has more values than the shapefile (which gives the coordinates for the districts). How can I create more districts to match the election results? And will this even solve my problem, or is there something else that I am missing? Also, the data is in spanish, but the relevant values Distritos=districts, partidos=political party, cargo=type of election, votos=votes.
library(maptools)
library(rgdal)
library(rgeos)
library(sf)
municipios <-readOGR("/Users/Desktop/Limite_partidos/Shapefile/Partidos.shp")
elec <- read.csv("/Users/Desktop/elections excel.csv", header = TRUE, stringsAsFactors = FALSE)
library(dplyr)
#merge datasets together
elec <-merge(elec, municipios, by=c("nam"), na.rm=TRUE)
plot(elec)
Link to election results (elec)
Link for shapefile (municipios)
First some sample data. This should be similar to what you get when
you read in your shapefile. The R spatial object is called a
SpatialPolygonsDataFrame. It carries a data.frame with the covariate
information about your polygons.
library(sp)
Sr1 <- Polygon(cbind(c(2,4,4,1,2),c(2,3,5,4,2)))
Sr2 <- Polygon(cbind(c(5,4,2,5),c(2,3,2,2)))
Sr3 <- Polygon(cbind(c(4,4,5,10,4),c(5,3,2,5,5)))
Sr4 <- Polygon(cbind(c(5,6,6,5,5),c(4,4,3,3,4)), hole = TRUE)
Srs1 <- Polygons(list(Sr1), "s1")
Srs2 <- Polygons(list(Sr2), "s2")
Srs3 <- Polygons(list(Sr3, Sr4), "s3/4")
SpP <- SpatialPolygons(list(Srs1,Srs2,Srs3), 1:3)
Spdf <- SpatialPolygonsDataFrame(SpP, data.frame(name = c("a", "b", "c"), row.names = c("s1", "s2", "s3/4")))
Now you have a spatial object that you can plot:
plot(Spdf)
and have a look at the attached data.frame of your spatial object. Here you need to have some identifier that you will match to your election results:
Spdf#data
You also have another dataframe with your 'election results' (Also with this identifier)
election <- data.frame(name = c("a", "c", "b"), voted = c(0.1, 0.2, 0.3))
Now match in the election results to you spatial object so you can plot it:
Spdf#data$voted <- election$voted[match(Spdf$name, election$name)]
To plot the polygons with the voted result as the colour of the polygon you need a pallette:
Spdf#data$colour <- heat.colors(3)[as.numeric(cut(Spdf#data$voted, 3))]
Then just plot:
plot(Spdf, col = Spdf#data$colour)
You can image that you will want to have more than 3 breaks in you
scale and you'll have more polygons but this is just an example. Good luck!
You can either select a single party (and then have a single value per administrative region) or facet by the party (and have as many small multiples as there are parties).
This really depends on what is your aim with the visualization. Given that there seems to be 48 results in your file the "small" multiples would be rather large, and filtering makes more sense.
For joining the shapefile and data frame with election results I suggest using one of the *_join functions from tidyverse package.
Consider this approach, built on assumption of filtering:
library(tidyverse)
library(sf)
tf_elec <- tempfile(fileext = ".csv") # create a temporary csv file
download.file("https://catalogo.datos.gba.gob.ar/dataset/1ae289f8-532c-4f69-a3c8-0268fe0ee390/resource/f8168491-4c38-4b03-82f1-b05fe43f8349/download/generales-2017.csv", tf_elec, quiet = T)
elec <- read_csv2(tf_elec) # read the data
tf_zip <- tempfile(fileext = ".zip") # a temoprary zip file
download.file("https://catalogo.datos.gba.gob.ar/dataset/627f65de-2510-4bf4-976b-16035828b5ae/resource/de607a34-b782-420f-93ed-35073a016e01/download/limite_partidos.zip", tf_zip, quiet = T)
unzip(tf_zip, files = 'Limite_partidos/GeoJSON/Partidos.geojson', exdir = tempdir(), junkpaths = T)
municipios <- st_read(paste0(tempdir(), '/Partidos.geojson'), quiet = T) # read the metro stations
src <- municipios %>%
left_join(elec, by = c('nam' = 'distrito')) %>%
filter(partido == 'VOTOS NULOS') #or what not... :)
ggplot() +
geom_sf(data = src, aes(fill = votos))
The most difficult part is downloading the data. The left_join() is in the second last chunk, and the very last is visualization - for the sake of simplicty I took the easy ggplot2 route, but also consider the excellent tmap package if you want your maps to really shine.
I am using similar to code to other scripts with different shape files from Statistics Canada. However, I can't get a simple script to work with a provincial map. I think the problem is simple but I can't see it.
setwd("D:\\OneDrive\\lfs_stuff")
project_folder<-getwd()
data_folder<-project_folder
library(tidyverse)
#now start the map
library(rgeos)
library(rgdal)
library(maptools)
library(sp)
library(mapproj)
library(ggplot2)
#get test data
mydata<-read_csv("map_data.csv",col_types=list(col_character(),col_double()))
print(mydata)
# shape file came from this link for a digital shape file
# http://www12.statcan.gc.ca/census-recensement/2011/geo/bound-limit/files-fichiers/2016/lpr_000a16a_e.zip
target_url<-"http://www12.statcan.gc.ca/census-recensement/2011/geo/bound-limit/files-fichiers/2016/lpr_000a16a_e.zip"
url_file<-"lpr_000a16a_e.zip"
download_target<-paste0(project_folder,"/",url_file)
download.file(target_url,download_target,mode="wb",quiet=FALSE)
unzip(download_target,overwrite=TRUE,exdir=data_folder)
provincial_shape_file<-gsub(".zip",".shp",download_target)
provincial_shp<-readOGR(dsn=provincial_shape_file,layer="lpr_000a16a_e")
#convert it to the reqired data structure. the id vbl will contain the provincial codes
prov_base_map<-fortify(provincial_shp,region="PRUID")
map_data_1<-merge(prov_base_map,as_data_frame(mydata),by="id")
map1<-ggplot()+
geom_map(data=map_data_1,map=map_data_1,stat="identity",
aes(map_id=id,x=long,y=lat,fill=(pch),group=group),
colour="black",size=0.3)+
coord_map()
print(map1)
The download for the shape file is in the script. The mydata file is shown below
"id","pch"
"10",0.667259786476859
"11",5.63186813186813
"12",2.12053571428572
"13",-0.563697857948142
"24",0.150669774230772
"35",1.15309092428315
"46",0.479282622139765
"47",1.70242950877815
"48",1.84482533036765
"59",1.96197656978394
Here's one way with sf (though I think the ultimate issue is not having the id being identified correctly):
library(sf)
library(httr)
library(tidyverse)
read.csv(text='"id","pch"
"10",0.667259786476859
"11",5.63186813186813
"12",2.12053571428572
"13",-0.563697857948142
"24",0.150669774230772
"35",1.15309092428315
"46",0.479282622139765
"47",1.70242950877815
"48",1.84482533036765
"59",1.96197656978394',
stringsAsFactors=FALSE,
colClasses = c("character", "double")) -> xdf
# cross-platform-friendly d/l with caching built-in
try(httr::GET(
url = "http://www12.statcan.gc.ca/census-recensement/2011/geo/bound-limit/files-fichiers/2016/lpr_000a16a_e.zip",
httr::write_disk("~/Data/lpr_00a16a_e.zip"),
httr::progress()
)) -> res
fils <- unzip("~/Data/lpr_00a16a_e.zip", exdir = "~/Data/lpr")
ca_map <- st_read(grep("shp$", fils, value=TRUE), stringsAsFactors = FALSE)
ca_map <- st_simplify(ca_map, TRUE, 10) # you don't need the coastlines to be that detailed
ca_map <- left_join(ca_map, xdf, by=c("PRUID"="id"))
ggplot(ca_map) +
geom_sf(aes(fill = pch)) +
viridis::scale_fill_viridis(direction=-1, option="magma") +
coord_sf()
As an aside, even though I simplified the shapefile (for faster plotting) I'd hunt around for a light[er]-weight GeoJSON version of the provinces since the one you grabbed has super fine-grained coastlines and you absolutely don't need that for a choropleth.
I'm trying to overlay some spatial data from a bigger SpatialPolygonsDataFrame (world size) to a smaller (country size), by doing these:
x <- c("rgdal", "dplyr",'ggplot2')
apply(x, library, character.only = TRUE)
est<-readOGR(dsn='/estados_2010',layer='estados_2010')
est_f<-fortify(est)
est$id<-row.names(est)
est_f<-left_join(est_f,est#data)
zon<-readOGR(dsn='/Zonas Homogeneas/gyga_ed_poly.shp',layer='gyga_ed_poly')
zon_f<-fortify(zon)
zon$id<-row.names(zon)
zon_f<-left_join(zon_f,zon#data)
t<-ggplot()+geom_polygon(data=zon_f,aes(x=long,y=lat,group=group,fill=GRID_CODE))
t+geom_polygon(data=est_f,aes(x=long,y=lat,group=group),fill=NA,color='red')+coord_fixed(xlim=est_f$long,ylim=est_f$lat,1)
Which is resulting in this:
I'm want to select only what is being plotted inside the polygon with the red lines.
If someone could help me with this issue, I'll appreciate
PS.: For those who want to reproduce the example completely by yourselves, the files are available in the links above to my google drive:
https://drive.google.com/open?id=0B6XKeXRlyyTDakx2cmJORlZqNUE
Thanks in advance.
Since you are using polygons to display the raster values, you can use a spatial selection via [ like in this reproducible example:
library(raster)
library(rgdal)
bra <- getData("GADM", country = "BRA", level = 1)
r <- getData("worldclim", res = 10, var = "bio")
r <- r[[1]]
r <- crop(r, bra)
r <- rasterToPolygons(r)
# bra and raster (now as polygons) have to have the same projection, thusly reproject!
bra <- spTransform(bra, CRSobj = proj4string(r))
here comes the magic!!
r <- r[bra, ]
let's look at the results:
library(ggplot2)
t <- ggplot()+
geom_polygon(data=r,aes(x=long,y=lat,group=group, fill = rep(r$bio1, each = 5)))
t +
geom_polygon(data=bra,aes(x=long,y=lat,group=group),fill=NA,color='red') + coord_map()