Finally gave up on my Rachio flow meter. Only OK while it lasted. Finally dug in and wrote some code to pull data from Flume and Rachio API to chart house usage vs. Zone. Result below:
ps: Yes - I have some zones that are permanently off. My main goal was to look for leaks and popped heads.
Update: one other representation.
that’s cool, were you able to automate this? I really think Rachio’s reporting of under/over is a more elegant solution, but these graphs are daaaaaamn 
Have it running in about 50 lines of R code (below). My problem is that my system is constantly popping and breaking heads, between construction and gardeners. And my Rachio meter has given up the ghost - never worked great. I don’t find the over / under very useful because it is always wrong.
Replaced hard-coded {user ID}, {client IDs}, {client secrets), {device ID}, and {auth tokens}. Toughest part was figuring out all the authorizations, and that the Flume {user ID} is the first numeric part of the {Client ID}.
library(tidyverse)
library(jsonlite)
library(curl)
library(httr)
# Set up data pull window
Start_text <- "2024-06-11 00:00:00"
End_text <- "2024-06-12 00:00:00"
#Auto conversions to POSIXct and epoch_milliseconds representations for the APIs
Start = as.POSIXct(Start_text)
End = as.POSIXct(End_text)
Start_epoch <- paste0(as.integer(Start) * 1000)
End_epoch <- paste0(as.integer(End) * 1000)
# Pull event history from Rachio
h <- new_handle(verbose = TRUE)
handle_setheaders(h,
"Content-Type" = "application/json",
"Authorization" = "Bearer {API key}")
request <- paste0("https://api.rach.io/1/public/device/{device ID}/event?startTime=", Start_epoch, "&endTime=", End_epoch)
con <- curl(request, handle = h)
# Format and filter Rachio history to just watering starts and last end
watering <- prettify(readLines(con)) %>%
fromJSON() %>%
mutate(eventDate = as.POSIXct(eventDate/1000)) %>%
mutate(Zone = str_extract(summary,"^.+(began|completed)"), Event = paste(Zone, subType)) %>%
filter(str_detect(subType, 'ZONE_STARTED') | str_detect(subType, 'ZONE_COMPLETED')) %>%
filter(str_detect(subType, 'ZONE_STARTED') | (str_detect(subType, 'ZONE_COMPLETED') & eventDate == max(eventDate))) %>%
select(eventDate, Zone, subType)
# Pull Minute Usage data from flume - limited to 1 day - 1440 data 1 minute data points
url <- "https://api.flumewater.com/users/{user ID}/devices/{device ID}/query"
payload <- "{\"queries\":[{\"units\":\"GALLONS\",\"request_id\":\"History\",\"bucket\":\"MIN\",\"since_datetime\":\"Start\",\"until_datetime\":\"End\"}]}" %>%
str_replace('Start', Start_text) %>% str_replace('End', End_text)
encode <- "json"
response <- VERB("POST", url, body = payload, add_headers('authorization' = 'Bearer {auth token}'), content_type("application/json"), accept("application/json"), encode = encode)
ldata <- content(response, "parsed")
# Convert list to dataframe
minWaterUsage<- do.call(rbind.data.frame, ldata$data[[1]]$History) %>%
rename(GalperMin = value) %>% mutate(datetime = as.POSIXct(datetime))
# Use Rachio Zone run data to annotate Flume minutes
intervals <- c(Start, watering$eventDate, End)
labels <- c("Off", watering$Zone[1:NROW(watering)-1], "Off")
minWaterUsage$Zone <- cut(minWaterUsage$datetime, intervals, labels = labels)
DaySummary <- minWaterUsage %>% group_by(Zone) %>% summarise(Gallons = sum(GalperMin), Minutes=n())
# Plot the overnight sprinkler run
minWaterUsage %>% ggplot(aes(x=datetime, y=GalperMin, fill=Zone)) + geom_col() + theme_bw() +
xlim(c(Start, End)) +
ggtitle("Flume Flow vs. Sprinkler Zones") + ylab("Gallons per Minute") + xlab("Time")
Thanks for sharing. The reverse engineering of API calls is the fun, tho, isn’t it? Especially when you eventually succeed 
Definitely fun when one succeeds 
Postman helped with Rachio, but didn’t have templates available for Flume.