Using the dataRetrieval Stats Service

October 4, 2016

(This article was first published on The USGS OWI blog , and kindly contributed to R-bloggers)


This script utilizes the new dataRetrieval package access to the USGS Statistics Web Service. We will be pulling daily mean data using the daily value service in readNWISdata, and using the stats service data to put it in the context of the site’s history. Here we are retrieving data for July 12th in the Upper Midwest, where a major storm system had recently passed through. You can modify this script to look at other areas and dates simply by modifying the states and objects.

To run this code, we recommend having either dataRetreival version 2.5.13 (currently the latest release on CRAN) or version 2.6.1 (currently the latest Github release).

Get the data

There are two separate dataRetrieval calls here — one to retrieve the daily discharge data, and one to retrieve the historical discharge statistics. Both calls are inside loops to split them into smaller pieces, to accomodate web service restrictions. The daily values service allows only single states as a filter, so we loop over the list of states. The stats service does not allow requests of more than ten sites, so the loop iterates by groups of ten site codes. Retrieving the data can take a few tens of seconds. Once we have both the daily value and statistics data, the two data frames are joined by site number via dplyr’s left_join function. We use a pipe to send the output of the join to na.omit() function. Then we add a column to the final data frame to hold the color value for each station.

#example stats service map, comparing real-time current discharge to history for each site
#reusable for other state(s)
#David Watkins June 2016


#pick state(s) and date
states <- c("WI","MN","ND","SD","IA") <- "2016-07-12"

#download each state individually
for(st in states){

  stDV <- renameNWISColumns(readNWISdata(service="dv",
                                       stateCd = st,
                                       startDate =,
                                       endDate =
  if(st != states[1]){ <- full_join(,stDV)
    sites <- full_join(sites, attr(stDV, "siteInfo"))
  } else { <- stDV
    sites <- attr(stDV, "siteInfo")

#retrieve stats data, dealing with 10 site limit to stat service requests
reqBks <- seq(1,nrow(sites),by=10)
statData <- data.frame()
for(i in reqBks) {
  getSites <- sites$site_no[i:(i+9)]
  currentSites <- readNWISstat(siteNumbers = getSites,
                               parameterCd = "00060", 
  statData <- rbind(statData,currentSites)

statData.storm <- statData[statData$month_nu == month( & 
                            statData$day_nu == day(,]

finalJoin <- left_join(,statData.storm)
finalJoin <- left_join(finalJoin,sites) 

finalJoin[,grep("_va",names(finalJoin))] <- sapply(finalJoin[,grep("_va",names(finalJoin))], function(x) as.numeric(x))

#remove sites without current data 
finalJoin <- finalJoin[!$Flow),] 

#classify current discharge values
finalJoin$class <- NA

finalJoin$class[finalJoin$Flow > finalJoin$p75_va] <- "navy"
finalJoin$class[finalJoin$Flow < finalJoin$p25_va] <- "red"

finalJoin$class[finalJoin$Flow > finalJoin$p25_va & 
                  finalJoin$Flow <= finalJoin$p50_va] <- "green"
finalJoin$class[finalJoin$Flow > finalJoin$p50_va &
                  finalJoin$Flow <= finalJoin$p75_va] <- "blue"

finalJoin$class[$class) & 
                  finalJoin$Flow > finalJoin$p50_va] <- "cyan"
finalJoin$class[$class) & 
                  finalJoin$Flow < finalJoin$p50_va] <- "yellow"

#take a look at the columns that we will plot later:
##   dec_lon_va dec_lat_va class
## 1  -92.09389   46.63333  navy
## 2  -91.59528   46.53778  navy
## 3  -90.96324   46.59439  navy
## 4  -90.59000   46.39472  navy
## 5  -90.69630   46.48661  navy
## 6  -90.90417   46.49722  navy

Make the static plot

The base map consists of two plots. The first makes the county lines with a gray background, and the second overlays the heavier state lines. After that we add the points for each stream gage, colored by the column we added to finalJoin. In the finishing details, grconvertXY is a handy function that converts your inputs from a normalized (0-1) coordinate system to the actual map coordinates, which allows the legend and scale to stay in the same relative location on different maps.

#convert states from postal codes to full names
states <- stateCdLookup(states, outputType = "fullName")
map('county',regions=states,fill=TRUE, col="gray87", lwd=0.5)
map('state',regions=states,fill=FALSE, lwd=2, add=TRUE)
       col=finalJoin$class, pch=19)
title(paste("Daily discharge value percentile rank\n",,line=1)
par(mar=c(5.1, 4.1, 4.1, 6), xpd=TRUE)

legend.colors <- c("cyan","yellow",
legend.names <- c("Q > P50*","Q < P50*",
                "Q < P25",
                "P25 < Q < P50","P50 < Q < P75",
                "Q > P75")

       pch=19,cex = 0.75,pt.cex = 1.2,
       col = legend.colors,
       ncol = 2)
map.scale(ratio=FALSE,cex = 0.75,
          grconvertY(.2, "npc"))
text("*Other percentiles not available for these sites", cex=0.75,
     y=grconvertY(-0.08, "npc"))


Make an interactive plot

Static maps are great for papers and presentations. When possible, interactive maps allow the reader more flexibility to examine the data. The R leaflet package makes it easy to create useful interactive maps:


finalJoin$popup <- with(finalJoin, paste("",station_nm,"
", "Measured Flow:",Flow,"ft3/s
", "25% historical:",p25_va,"ft3/s
", "50% historical:",p50_va,"ft3/s
", "75% historical:",p75_va,"ft3/s")) leafMapStat <- leaflet(data=finalJoin) %>% addProviderTiles("CartoDB.Positron") %>% addCircleMarkers(~dec_lon_va,~dec_lat_va, color = ~class, radius=3, stroke=FALSE, fillOpacity = 0.8, opacity = 0.8, popup=~popup) leafMapStat <- addLegend(leafMapStat, position = 'bottomleft', colors= legend.colors, labels= legend.names, opacity = 0.8)

Disclaimer: The NWIS stats web service that dataRetrievalaccesses here is in beta, and its output could change in the future.


Please direct any questions or comments on dataRetrieval to:

To leave a comment for the author, please follow the link and comment on their blog: The USGS OWI blog . offers daily e-mail updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...

If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.


Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)