In impact-initiatives-geospatial/surveyGEER: Bringing relevant GEE Data Into Your Assessment

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(surveyGEER)
library(rgee)
library(sf)
library(tidyverse)
library(tidyrgee)
ee_Initialize()

Sentinel 5 P

• has offline versions and NRT versions
• NRT are ingested more quickly and hence more up to date.
• We comparing NRT vs Offline NO2 we NRT has images from yesterday whereas the offline version appears to have been last updated approximately 10 days ago

# Sentinel-5P OFFL SO2: Offline Sulphur Dioxide -1113.2 meters
so2 <- ee$ImageCollection("COPERNICUS/S5P/OFFL/L3_SO2")
# Sentinel-5P OFFL NO2: Offline Nitrogen Dioxide
no2_off <- ee$ImageCollection("COPERNICUS/S5P/OFFL/L3_NO2")
no2_nrt <- ee$ImageCollection("COPERNICUS/S5P/NRTI/L3_NO2")

# Sentinel-5P OFFL CH4: Offline Methane
ch4 <- ee$ImageCollection("COPERNICUS/S5P/OFFL/L3_CH4")

no2_off_tidy <-  as_tidyee(no2_off)
no2_nrt_tidy <-  as_tidyee(no2_nrt)

no2_off_tidy$vrt$time_start |> max()
no2_nrt_tidy$vrt$time_start |> max()

I think that for most of my current purposes and this demonstration the offline version is fine.

From reading it seems that this data can be noisy... therefore it makes sense to aggregate... will also have to deal with clouds

bbox_ee <- col_pt_data_clean |> 
  st_bbox() |> 
  st_as_sfc() |> 
  sf_as_ee()

cloud_mask_s5p <-  function(image,threshold) {
      return(
        image$updateMask(image$select("cloud_fraction")$lt(threshold))
      )
    }

# // Map the cloud masking functions over NO2 data
no2_cloudmasked = no2_off$
  select(list("NO2_column_number_density",
              "tropospheric_NO2_column_number_density",
              'cloud_fraction'))$
  filterDate('2018-05-01', '2022-05-01')$
  # filterBounds(geometry)
  map(function(img){
    cloud_mask_s5p(img,threshold = 0.05)}
    )


NO2_max = 0.0010;
NO2_min = 0;
NO2_viz = list(
  min= NO2_min,
  max= NO2_max,
  opacity= 0.5,
  palette= c("black", "blue", "purple", "cyan", "green", "yellow", "red"),
  bands="NO2_column_number_density"
  )

Map$centerObject(bbox_ee,7)
Map$addLayer(no2_cloudmasked$median(),NO2_viz,"NO2")

so2$first()$bandNames()$getInfo()

 # // Map the cloud masking functions over NO2 data
so2_cloudmasked = so2$
  select(list("SO2_column_number_density",
              "SO2_column_number_density_amf",
              "SO2_column_number_density_15km",
              'cloud_fraction'))$
  filterDate('2018-05-01', '2022-05-01')$
  # filterBounds(geometry)
  map(function(img){
    cloud_mask_s5p(img,threshold = 0.05)}
    )




NO2_max = 0.0001;
NO2_min = 0;
NO2_viz = list(
  min= NO2_min,
  max= NO2_max,
  opacity= 0.5,
  palette= c("black", "blue", "purple", "cyan", "green", "yellow", "red"),
  bands="SO2_column_number_density"
  )

Map$centerObject(bbox_ee,7)
Map$addLayer(so2_cloudmasked$median(),NO2_viz,"NO2")

swims<- read_csv("../REACH_SOM/SWIMS_LiveMap_Dataset_20220808-135749.csv")
reach_hc <-  readxl::read_excel("../REACH_SOM/Master_Source_16062022.xlsx","Non_IDP")
reach_new_idp <-  readxl::read_excel("../REACH_SOM/Master_Source_16062022.xlsx","new_IDP_sites")
reach_old_idp <-  readxl::read_excel("../REACH_SOM/Master_Source_16062022.xlsx","old_IDP_sites")

con_som <- quick_db_con(schema = "som")
con <- quick_db_con(schema = "public")
swims |>
  glimpse()
swims |> count(water_source_type,functioning) |> print(n=21)
swims$functioning
swims_sf <- swims |> 
  mutate(
    type_functioning= case_when(
      is.na(functioning)~ glue::glue("{water_source_type} (no data)"),
      functioning == "No"~ glue::glue("{water_source_type} (non-functional)"),
      functioning == "Yes"~ glue::glue("{water_source_type} (functional)"),
      functioning == "Abandoned"~ glue::glue("{water_source_type} (Abandoned)"),
    )
  ) |> 
  st_as_sf(coords= c("longitude","latitude"), crs= 4326)


swims_utm <- swims_sf |> 
  reach_reproject_utm("som")

swims_functional_water_points <-  swims_utm |> 
  filter(water_source_type!="Dam", functioning=="Yes",!is.na(functioning))


con_som <- quick_db_con(schema = "som") 
buildings <- st_read(con_som,"som_bay_buildings_google")
building_pts <- buildings |> 
  st_centroid()
building_pts_utm <- building_pts|> 
  reach_reproject_utm("som")

dist_to_closest_waterpoint <- st_nearest_points(x =building_pts_utm,y = swims_functional_water_points )
nearest_water_pt_idx <- dist_to_closest_waterpoint
dist_to_closest_waterpoint<-st_distance(building_pts_utm,swims_functional_water_points[nearest_water_pt_idx,])

building_pts_utm |> 
  mutate(dist_to_wp = dist_to_closest_waterpoint)







swims |> 
  filter(is.na(longitude))
swims_sf |> 
  filter(is.na(longitude))
swims_sf |> 
  st_write(con,"som_swims_water_points_2019")
length(unique(swims$code))
swims |> 
  count(water_source_type)

reach_hc |> count()
reach_hc |> count(Access)
reach_hc |> count(NAME)

reach_hc |> filter(is.na(MGRS_CODE_X))
reach_hc |> filter(is.na(MGRS_CODE_Y)|is.na(MGRS_CODE_X))


reach_hc |> 
  mutate(missing_cooridnate= if_else(is.na(MGRS_CODE_X)|is.na(MGRS_CODE_Y),
                                     "missing coordinates" ,
                                     "has coordinates")) |> 
  janitor::tabyl(missing_cooridnate)


reach_new_idp |> 
  glimpse()

reach_old_idp |> 
  glimpse()

reach_new_idp |> 
  mutate(missing_cooridnate= if_else(is.na(gps_latitude)|is.na(gps_longitude),
                                     "missing coordinates" ,
                                     "has coordinates")) |> 
  janitor::tabyl(missing_cooridnate)

reach_old_idp |> 
  mutate(missing_cooridnate= if_else(is.na(gps_latitude)|is.na(gps_longitude),
                                     "missing coordinates" ,
                                     "has coordinates")) |> 
  janitor::tabyl(missing_cooridnate)

bay_fc <- ee$FeatureCollection("FAO/GAUL/2015/level1")$
  filter(ee$Filter$eq('ADM0_NAME', 'Somalia'))$
  filter(ee$Filter$eq('ADM1_NAME', 'Bay'))
google_buildings <-  ee$FeatureCollection('GOOGLE/Research/open-buildings/v1/polygons')

buildings_bay = google_buildings$filter('confidence >= 0.60')$filterBounds(bay_fc)
# buildings_bay_fc <- buildings_bay |> rgee::ee_as_sf(via = "drive")
# buildings_bay_sf <- buildings_bay_fc




task_vector<- ee_table_to_drive(collection = buildings_bay,fileFormat = "GeoJSON")
task_vector$start()

ee_monitoring(task_vector) # optional
# amk_geom <- ee_gcs_to_local(task = task_vector)
buildings_bay_local <- ee_drive_to_local(task = task_vector)

impact-initiatives-geospatial/surveyGEER documentation built on Feb. 4, 2023, 12:13 p.m.