In bmkor/hktraffic: Display HK Traffic Data

Display Hong Kong Traffic Data

In this small exercise, we'll try to parse the real-time traffic speed data on 6 hundred-something links(or roads) in Hong Kong, covering Tuen Mum (TM), Shatin (ST), Kowloon (K) and Hong Kong island (HK) and display the data on the OpenStreetMap by package leaflet. Each link is either belonged to Major Route (speed limit >= 70 km/h) or Urban Road (speed limit = 50 km/h)

Disclaimer

This exercise contains information that is copied or extracted from data made available by the Government of Hong Kong Special Administrative Region (the "Government") at https://DATA.GOV.HK/ ("DATA.GOV.HK"). The provision of information copied or extracted from or a link to DATA.GOV.HK at this website or in relation to the product or service shall not constitute any form of co-operation or affiliation by the Government with any person in relation to this website, the product or the service or any contents herein. Nothing in this website, the product or the service shall give rise to any representation, warranty or implication that the Government agrees with, approves of, recommends or endorses any contents of this website or the product or the service. The Government does not have any liability, obligation or responsibility whatsoever for any loss, destruction or damage (including without limitation consequential loss, destruction or damage) howsoever arising from or in respect of your use or misuse of or reliance on or inability to use any contents herein.

1. Grep and parse the specificaton of the links(or roads) in Hong Kong Traffic Data

The real-time traffic speed data can be downloaded through this link http://resource.data.one.gov.hk/td/speedmap.xml. And the specification of the link can be downloaded via this link http://static.data.gov.hk/td/traffic-speed-map/en/tsm_dataspec.pdf.

We first build a function parsing the specification pdf and return to a SpatialLinesDataFrame. Note that the coordinates in the pdf are in HK1980Grid which corresponds to epsg:2326.

require(pdftools)
require(sp)

getRoads<-function(specURL="http://static.data.gov.hk/td/traffic-speed-map/en/tsm_dataspec.pdf"){
  tmp<-tempfile()
  roadMetaData<-specURL
  download.file(roadMetaData,dest=tmp,quiet =T)

  roadMetaData<-pdftools::pdf_text(tmp)
  roadMetaData<-roadMetaData[-1]

  res<-sapply(roadMetaData, function(x){
    stmp<-unlist(strsplit(x,"\\n"))[-(1:2)]
    gsub("^\\s","",gsub(" ROAD", "_ROAD",gsub(" ROUTE", "_ROUTE",gsub("\\s+"," ",stmp))))
  })
  names(res)<-1:length(res)
  res<-c(paste("ROUTE","START","START_E","START_N","END","END_E","END_N","DISTRICT","ROUTE_TYPE",sep=" "),as.vector(unlist(res)))
  write(res,tmp)
  roadMetaData<-read.csv(tmp,sep=" ",stringsAsFactors = TRUE)
  rownames(roadMetaData) <- roadMetaData$ROUTE

  if(require("sp",character.only = T, quietly = T)){
    roads<-apply(roadMetaData,1,function(r){
      m<-matrix(as.numeric(c(r[3],r[4],
                             r[6],r[7])), nrow = 2, byrow =T)
      m<-data.frame(m)
      sp::coordinates(m) <- ~X1+X2
      sp::SpatialLines(list(sp::Lines(list(sp::Line(m)), r[1])),sp::CRS("+init=epsg:2326"))
    })
    roads<-do.call(rbind,roads)
    roads<-sp::SpatialLinesDataFrame(roads,roadMetaData[,c("ROUTE","DISTRICT","ROUTE_TYPE")])
    roads<-sp::spTransform(roads,sp::CRS("+init=epsg:4326"))
    roads
  }else{
    NA
  }
}

summary(getRoads())

Note that more links are of Major route type and Shatin got the highest number of roads/links.

2. Plotting the roads out

We try to plot out those links/roads by leaflet.

```r require(leaflet)

plotRoads<-function(roads=getRoads()){ if(require("leaflet",character.only = T,quietly =T)){ roads$dashPattern<-as.character(c(1,c("15, 10, 5"))[as.numeric(roads$ROUTE_TYPE)]) pale<-c("yellow","orange","blue","brown") dPal<-leaflet::colorFactor(pale,domain=roads$DISTRICT) leaflet(data=roads) %>% addTiles(group = "OSM (default)") %>% addProviderTiles(providers$Stamen.Toner, group = "Toner") %>% addProviderTiles(providers$Stamen.TonerLite, group = "Toner Lite") %>% addPolylines(color=~dPal(DISTRICT),dashArray=~dashPattern, popup=~paste0("route type: ",ROUTE_TYPE), label = ~ROUTE) %>% addLegend(position="bottomright", pal = dPal, values = ~DISTRICT, title = "District", labels=~DISTRICT) %>% addLayersControl( baseGroups = c("OSM (default)", "Toner", "Toner Lite"), options = layersControlOptions(collapsed = FALSE) ) } }

plotRoads()

Note that the roads/links are straight lines which are not realistic and Urban Roads are in dashed lines.

##3. Plotting the current HK Traffic

Finally, we get the real-time traffic data (in XML format) through the link <http://resource.data.one.gov.hk/td/speedmap.xml> and highlight the Spatial lines (or roads or links) for the traffic satuartion and etc.

````r
require(XML)
plotCurrentTraffic<-function(trafficURL="http://resource.data.one.gov.hk/td/speedmap.xml", roads=getRoads()){
  if(require("leaflet",character.only = T,quietly =T) & require("XML",character.only = T,quietly =T)){
    tdata<-XML::xmlParse(trafficURL)
    tdata<-XML::xmlToDataFrame(tdata, stringsAsFactors = F)
    tdata$ROAD_SATURATION_LEVEL<-gsub("TRAFFIC ","",tdata$ROAD_SATURATION_LEVEL)
    tdata$ROAD_SATURATION_LEVEL <- factor(tdata$ROAD_SATURATION_LEVEL)
    tdata$CAPTURE_DATE<-strptime(tdata$CAPTURE_DATE,format="%Y-%m-%dT%H:%M:%S")

    pale<-c("green","yellow","red")
    names(pale)<-c("GOOD","AVERAGE","BAD")
    pale<-as.vector(pale[levels(tdata$ROAD_SATURATION_LEVEL)])

    pal<-leaflet::colorFactor(pale,domain=tdata$ROAD_SATURATION_LEVEL)

    roads$dashPattern<-as.character(c(1,c("15, 10, 5"))[as.numeric(roads$ROUTE_TYPE)])
    roads<-subset(roads, ROUTE %in% tdata$LINK_ID)
    roads@data<-cbind(ROUTE=roads$ROUTE,dashPattern=roads$dashPattern,
                      tdata[match(as.vector(roads$ROUTE),tdata$LINK_ID),])
    lastCaptureTime<-max(unique(tdata$CAPTURE_DATE))
    lastCaptureTime<-strftime(lastCaptureTime,format="%H:%M, %d %b %Y")
    leaflet(data=roads) %>%
      addTiles(group = "OSM (default)") %>%
      addProviderTiles(providers$Stamen.Toner, group = "Toner") %>%
      addProviderTiles(providers$Stamen.TonerLite, group = "Toner Lite") %>%
      addPolylines(color=~pal(ROAD_SATURATION_LEVEL),dashArray=~dashPattern,
                   popup=~paste0("average speed: ",TRAFFIC_SPEED, " km/h"), label = ~ROUTE) %>%
      addLegend(position="bottomright", pal = pal,
                values = ~ROAD_SATURATION_LEVEL, title = paste0("Road Saturation Level at ",lastCaptureTime),
                labels=~ROAD_SATURATION_LEVEL) %>%
      addLayersControl(
        baseGroups = c("OSM (default)", "Toner", "Toner Lite"),
        options = layersControlOptions(collapsed = FALSE)
      )  }
}

plotCurrentTraffic()

Note that the time stamp is the latest time.

4.(Optional) ~~Download from github~~

~~If you need the above functions for whatever reasons, you may just install this package hktraffic from my github.~~ ~~devtools::install_github("bmkor/hktraffic")~~

Outdated. I'll update this little package later...

5. ESRI

Found esri collabarated with HK goverenment and happy to see that they got a set of road polylines for the TD traffic map.

So, let's try to plot those lines on the map.

First, we create a handy function to load esri data

getRoadsFromESRI<-function(esriURL = "https://opendata.arcgis.com/datasets/91b65a5464d9403687be340146a18f2a_0.geojson"){
  require(geojsonio,quietly = T)
  tmp<-geojson_read(esriURL, what="sp")
  names(tmp)[names(tmp) == "LINK_ID"] <- "ROUTE"
  names(tmp)[names(tmp) == "ROAD_TYPE"] <- "ROUTE_TYPE"
  tmp<-subset(tmp, !is.na(ROUTE_TYPE)) # some roads with NA route type
  tmp
}

Then, we plot out the lines...

plotCurrentTraffic(roads=getRoadsFromESRI())

Please zoom in to see how the polylines are roads.

Interesting that the some polylines (e.g. Island Eastern Corridor:877-46498 ) are not on the roads. This is because esri used HERE map whereas we are using OpenStreetMap.

Unfortunately, HERE map requires subscription and I'm not going to subscribe it. Perhaps I can try to remap those polylines onto the roads of the OpenStreetMap later...

bmkor/hktraffic documentation built on May 12, 2019, 10:27 p.m.