Can't plot a R interactive map comtainning my shapefile and my raster data in power BI

General
1

Hello!
I'm trying to load my script in power BI, which is:

library(ggplot2)
library(ggmap)
library(plotly)
library(raster)
library(sf)
library(rgdal)
library(rgeos)
library(mapview)

setwd("C:/Users/Daniel/Desktop/IDGeo/Earth analytics course_Learn data science/data/week 4")

ndwi <- raster("california/SJER/vector_data/2017_06_04_NDWI_CLIP.tif")

pivos <- readOGR(dsn = "california/SJER/vector_data",
"aneis_anapaula_0904")

M2<- mapView(ndwi, alpha.regions = 0.50, na.color = "transparent")
m4 <- mapview(pivos, lwd = 1.5, color = "gray40")

M2 + m4

ndwi is my raster data
pivos is my shapefile

In R it works perfectly and returns an interactive map containing my raster data plus myshapefile.

Anyone knows what could be happening?
Or, anyone could suggest another way to plot raster + shapefile interactive maps in power BI using R?
Thanks very much, best regards.

1 Like
2

I believe that {mapview} is not compatible with Power BI.

Are you absolutely positively determined on using Power BI? Your piece of code looks rather static, and could possibly be rendered as html via leaflet.

1 Like
3

Hello! Yes, I need to use Power BI...
Do you have any idea? To plot a raster + shapefile, allowing to zoom them?
Thanks very much!

4

What I have done in the past was to create a TopoJSON (not very difficult in {sf} workflow) out of the shapefile and then used it to create a choropleth map. There is a walkthrough to it on the Power BI pages.

It might be something along these lines (note that your example is not quite reproducible, so I can not make sure).

yer_shapefile <- sf::st_read("yer_shapefile.shp") # you will want to be more specific :)
yer_raster <- raster:.raster("your_raster.tif") %>% # original raster
   sf:.st_as_sf() # grid - without values - as a sf object

geojsonio::topojson_write(yer_r_object, # the sf object - first raster, then shapefile
                          file = "whatever.topojson", # or what not...
                          object_name = "whatever") # or what not...

In Power BI you would need to re-link the data to empty scaffolding created from the topojson.

This should work, but it is kinda hassle, IMHO.

An alternative would be generating the plot in R via {leaflet} and "plotting" it from Power BI as a chunk of HTML.

2 Likes
5

Hi, thanks for your attention! Well, a choropleth map won't work for what I need to demonstrate...
About the HTML chunk, I think it could work..
Should I generate a Rmardown file and input it in power BI ?
Again, thanks very much!

6

It need not be a rmd file: consider this example (drawing heavily on my earlier answer to a related question).

The last line is what matters, the rest is just a piece of reproducible leaflet

Once you have the html file saved you can embed the code into Power BI (there should be a ton of walkthroughs for that around...)

library(tidyverse)
library(leaflet)
library(tmap) # for the Metro dataset
library(sf)

data("metro") # from tmap

metro <- metro %>% 
  select(pop = pop2020) %>% 
  mutate(pop = pop / max(pop)) # normalise the values

mapa <- st_read("./data/maps/ne_50m_admin_0_countries.shp") # replace with your path :)

pal <- colorNumeric(palette = "Reds",  domain = mapa$GDP_MD_EST)

asdf <- 
  leaflet() %>%
  addProviderTiles(providers$CartoDB.Positron) %>%
  setView(lng = 14.46, lat = 50.07, zoom = 1) %>%
  addPolygons(data = mapa, 
              fillColor = ~pal(mapa$GDP_MD_EST),
              fillOpacity = 0.75,
              stroke = F) %>% 
  addCircleMarkers(data = metro,
                   radius = 20 * metro$pop,
                   stroke = F)

htmlwidgets::saveWidget(asdf, "asdf.html") # put something more specific than asdf here :)
2 Likes
7

Hello! Thanks very much for your script..
I'm quite beginner in R and I don't know exactly where should I put my raster data in this script...
should I add another object using the raster library?

Again,
thank you!

8

Try this code; as I don't have access to your filesystem I am coding blind. So bear with me if there is a missing pipe or the result needs some tweaking to make colors right or what not... :slight_smile:

library(tidyverse)
library(htmlwidgets)
library(raster)
library(sf)

# these are your files from code above, so I assume them to work...
pivos <- st_read("california/SJER/vector_data/aneis_anapaula_0904.shp")
ndwi  <- raster("california/SJER/vector_data/2017_06_04_NDWI_CLIP.tif")

chrt <- leaflet() %>%
  addProviderTiles(providers$CartoDB.Positron) %>% # a basemap; not required but nice...
  addPolygons(data = pivos, 
              fillOpacity = 0,
              stroke = T) %>%
  addRasterImage(ndwi)

saveWidget(chrt, "chart.html")
1 Like
9

Hi, thanks again and again... however, only the basemap is plotted... The raster and shp don't load =(
I received those waring messages only:
Warning messages:
1: sf layer is not long-lat data
2: sf layer has inconsistent datum (+proj=utm +zone=23 +south +ellps=GRS80 +units=m +no_defs).
Need '+proj=longlat +datum=WGS84'

I've checked the crs and it is like:

crs(pivos)
[1] "+proj=utm +zone=23 +south +ellps=GRS80 +units=m +no_defs"
crs(ndwi)
CRS arguments:
+proj=utm +zone=23 +south +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs

Both are in the same CRS, do you have any idea about wjat could be happening?

Thanks very much, you are very kind!!!

10

Sorry, I can't really help you without the files in question :frowning:

The best I can think of is

pivos <- st_read("california/SJER/vector_data/aneis_anapaula_0904.shp") %>%
  st_transform(4326) # i.e. WGS84
1 Like
11

Hi @jlacko ! Now the HTML is perfectly working!
I'm not able to display it in power BI yet, but I will find a way (if you have any, please suggest, haha) .
Anyway, you solved my firts problem, thank you!

12

Glad to be of service! :slight_smile:

13

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.