是否可以通过R中的多个组进行汇总？

Question

I have a camera trap dataset with Filenames, SiteID, Species, Count, Date, Time, etc. What I am trying to do is create a record table where I have the MAXIMUM number of independent detections, for each species detected, for each camera station (ie if the independent interval was set at 30 mins, and there was a detection of 2 deer and a detection of 13 deer within the same 30 minute interval, I want the 13 to be used instead of the 2).

Original data:

File     SiteID     Date            Time       Species     Count
Can_001  YVR01      03-May-2018     21:34:25   Squirrel    3
Can_001  YVR01      03-May-2018     21:34:58   Squirrel    3
Can_001  YVR01      03-May-2018     21:36:25   Squirrel    1

What I have done so far is to try and first group by siteID, then by Species, then by Date, and then to create a column where there are 30 minute time intervals, from which I need to then figure out how to get the maximum 'Count' value within the time interval - these will be the detections I am using.

species_group <- group_by(y4, SiteID) %>% group_by(Species) %>% group_by(Date) %>% group_by(Interval_Time=floor_date(DateTimeOriginalp, "30 minutes"))

I was able to get to the stage where the 30 minute interval period was created and column was created, but after this point no summarise(), aggregate(), tapply() etc function seems to work, as it won't allow me to pull up the "Interval_Time" column created. The new Interval_Time column is in dttm format, and shows up when I view and call the species_group dataframe. What I need to do now is get the MAX count of each species within these intervals. This is what I tried (ie outside of the pipe):

speciesgroup3 <- aggregate(species_group$Count, by=list(species_group$Interval_Time), max)

Which returned a dataframe of just two columns, the maximum count and the Interval_Time...which isn't useful as I need this data separated first by site and then by species.

For the life of me I can't figure out why I can't call Interval_Time as a column within the pipe above. Any help would be greatly appreciated!

Answer 1

Maybe you could use:

aggregate(. ~Interval_Time+SiteID, data=species_group, max, na.rm=TRUE)

Look for instance here for a similar problem https://stats.stackexchange.com/questions/169056/aggregate-all-data-by-date-and-id

Answer 2

This should be close to what you are looking for, using the dplyr functions included in library tidyverse

library(tidyverse)
library(lubridate)
df = read.table(text="
File     SiteID     Date            Time       Species     Count
Can_001  YVR01      03-May-2018     21:34:25   Squirrel    3
Can_001  YVR01      03-May-2018     21:34:58   Squirrel    3
Can_001  YVR01      03-May-2018     22:01:25   Squirrel    1
Can_001  YVR01      03-May-2018     21:34:58   Deer        5
Can_001  YVR01      03-May-2018     21:36:25   Deer        7
", header=T)

# Use mutate and the lubridate::mdy_hms to derive a proper date
# column from the text date time 
df2 <- df %>%
  mutate(DateTime = mdy_hms(paste(Date, Time)),
         period = floor_date(DateTime, "30 mins")) %>%
         select(-Date, -Time)

# File SiteID  Species Count            DateTime              period
# 1 Can_001  YVR01 Squirrel     3 2018-03-20 21:34:25 2018-03-20 21:30:00
# 2 Can_001  YVR01 Squirrel     3 2018-03-20 21:34:58 2018-03-20 21:30:00
# 3 Can_001  YVR01 Squirrel     1 2018-03-20 22:01:25 2018-03-20 22:00:00
# 4 Can_001  YVR01     Deer     5 2018-03-20 21:34:58 2018-03-20 21:30:00
# 5 Can_001  YVR01     Deer     7 2018-03-20 21:36:25 2018-03-20 21:30:00

# Summarize dow to the period level, applying the max function within the group by
df2 %>% 
  group_by(SiteID, Species, period) %>%
  summarize(n = max(Count))

# Groups:   SiteID, Species [?]
# SiteID Species  period                  n
# <fct>  <fct>    <dttm>              <dbl>
# 1 YVR01  Deer     2018-03-20 21:30:00     7
# 2 YVR01  Squirrel 2018-03-20 21:30:00     3
# 3 YVR01  Squirrel 2018-03-20 22:00:00     1