library(tidyverse)
library(lubridate)
library(openair)
library(threadr)
library(leaflet)

# store the filename in a variable
filename <- "https://data.rivm.nl/data/luchtmeetnet/Vastgesteld-jaar/2019/2019_NO2.csv"

mydata <- read.table(file = filename,
                     header = TRUE,
                     sep = ";",
                     dec = ".",
                     comment.char = "#",
                     quote = "\"") |> 
  dplyr::mutate(date = as.POSIXct(begindatumtijd, 
                                  format = "%Y-%m-%dT%H:%M:%S",
                                  tz="Etc/GMT-1"),
                date_end = as.POSIXct(einddatumtijd, 
                                      format = "%Y-%m-%dT%H:%M:%S",
                                      tz="Etc/GMT-1")-1) |>
  threadr::time_pad(interval = "hour",
                    by = "meetlocatie_id")

# show contents of the variable
head(mydata)

# Adding other data is easy, for example from a different matrix

# set the filename
filename <- "https://data.rivm.nl/data/luchtmeetnet/Vastgesteld-jaar/2019/2019_groep_HC_aerosol.csv"

mydata2 <- read.table(file = filename,
                     header = TRUE,
                     sep = ";",
                     dec = ".",
                     comment.char = "#",
                     quote = "\"") |> 
  dplyr::mutate(date = as.POSIXct(begindatumtijd, 
                                  format = "%Y-%m-%dT%H:%M:%S",
                                  tz="Etc/GMT-1"),
                date_end = as.POSIXct(einddatumtijd, 
                                      format = "%Y-%m-%dT%H:%M:%S",
                                      tz="Etc/GMT-1")-1) |>
  threadr::time_pad(interval = "day",
                    by = c("meetlocatie_id",
                           "component"))


# combine the data into one file is easy, as all files have the same structure:
mydata <- dplyr::bind_rows(mydata,
                           mydata2)

# cleanup some of the column names, so we can easily use other packages to plot
# the data
mydata <- mydata |>
  dplyr::rename(site = meetlocatie_id,
                variable = component,
                value = waarde)

# Since the data is now in long, we can use the R package threadr to plot the
# time variation as a ggplot:
threadr::plot_time_variation(df = mydata |> 
                               filter(site == "NL10641",
                                      variable == "NO2"),
                             colours = "steelblue",
                             legend_name = openair::quickText("NO2"))

# ggplot is really nice, as it allows for modification after creating the 
# initial plot. However, if we want to use openair, then we need to do a minor 
# modification to the data structure:

mydata_openair <- mydata |>
  tidyr::pivot_wider(names_from = variable,
                     values_from = value)

openair::timeVariation(mydata = mydata_openair |> filter(site == "NL10641"),
                       pollutant = "NO2")

# Lets add some meta data

# set the filename
filename <- "https://data.rivm.nl/data/luchtmeetnet/luchtmeetnet_meetlocaties.csv"

meta_sites <- read.table(file = filename,
                         header = TRUE,
                         sep = ";",
                         dec = ".",
                         comment.char = "#",
                         quote = "\"") |> 
  dplyr::mutate(begindatum_locatie = as.POSIXct(begindatum_locatie, 
                                            format = "%Y-%m-%dT%H:%M:%S",
                                            tz="Etc/GMT-1"),
                einddatum_locatie = as.POSIXct(einddatum_locatie, 
                                           format = "%Y-%m-%dT%H:%M:%S",
                                           tz="Etc/GMT-1")-1)

# show what we have
head(meta_sites)

# We changed the names of the columns in the mydata, so we need to do the same
# here. And in this case we are only interested in the coordinates.

meta_sites <- meta_sites |>
  dplyr::rename(site = meetlocatie_id,
                lat = breedtegraad,
                lon = lengtegraad) %>% 
  dplyr::select(site, lat, lon)

# Join the two together, using a left_join to retain all the measurements:

mydata <- dplyr::left_join(mydata,
                           meta_sites,
                           by = join_by(site))


# plot all the sites in the data frame using leaflet
leaflet::leaflet(data = mydata |>
                   select(site, lat, lon) |>
                   distinct()) |> 
  leaflet::addTiles() |>
  leaflet::addMarkers(~lon, ~lat, popup = ~site, label = ~site)