R help - Mar 2009 - testing column data against criteria, point by point

I am fairly new to R and I would like to do the following, but do not know where
to start. Any help or direction would be appreciated.
I have a time series of snow depth measurements. I would like to determine the
depth of snowfall for each snowfall event. There is noise in the data so I only
want to add data values if the subsequent depth is greater than the previous by
a certain margin. I am only interested in calculating snow accumulation events.
Example data:
 
Time	 depth	
1	 84.3	
2	 84.5	
3	 86	
4	 86.1	
5	 85.8	
6	 86.7	
7	 87.9	
8	 89.1	
9	 90	
10	 89	
11	 88	
12	 88	
13	 89.1	
14	 90	
15	 91.2	
16	 89.9	
...	 ...	
I would like to create a second data frame from the data that looks something
like this:
 
Event	 InitialDepth	 FinalDepth	 Accumulation	 InitialTime	 FinalTime	
1	 84.3	 90	 5.7	 1	 9	
2	 88	 91.2	 3.2	 11	 15	
...	 					
 
I would like to write a program that progresses through the depth column, point
by point, to test if (i+1) - i > x. (where I will set x to exlude the noise
in the data). As long as i+1 is greater than or equal to i, then the initial
depth stays at the first data point and the final value changes to that in i+n.
Once the test is false, this indicates the end of the event, the accumulation is
calculated, all values are saved as event X and a new event is started.
I tried using ifelse(), but I do not know how to move through the data and then
save the initial and final values and time stamps in another table.
 
Thank you very much for your time.
 
Kara

Hi


r-help-bounces at r-project.org napsal dne 05.03.2009 09:46:17:
> I am fairly new to R and I would like to do the following, but do not 
know 
> where to start. Any help or direction would be appreciated.
> I have a time series of snow depth measurements. I would like to 
determine the
> depth of snowfall for each snowfall event. There is noise in the data so 
I 
> only want to add data values if the subsequent depth is greater than the 
> previous by a certain margin. I am only interested in calculating snow 
> accumulation events.
> Example data:
> 
> Time    depth 
> 1    84.3 
> 2    84.5 
> 3    86 
> 4    86.1 
> 5    85.8 
> 6    86.7 
> 7    87.9 
> 8    89.1 
> 9    90 
> 10    89 
> 11    88 
> 12    88 
> 13    89.1 
> 14    90 
> 15    91.2 
> 16    89.9 
> ...    ... 
> I would like to create a second data frame from the data that looks 
something like this:
> 
> Event    InitialDepth    FinalDepth    Accumulation    InitialTime 
FinalTime 
> 1    84.3    90    5.7    1    9 
> 2    88    91.2    3.2    11    15 
> ... 
> 
> I would like to write a program that progresses through the depth 
column, 
> point by point, to test if (i+1) - i > x. (where I will set x to exlude 
the 
> noise in the data). As long as i+1 is greater than or equal to i, then 
the 
> initial depth stays at the first data point and the final value changes 
to 
> that in i+n. Once the test is false, this indicates the end of the 
event, the 
> accumulation is calculated, all values are saved as event X and a new 
event isstarted. 
> I tried using ifelse(), but I do not know how to move through the data 
and 
> then save the initial and final values and time stamps in another table.
It is probably possible but I would use rle and cumsum

# which of the data are increasing
ind<-rle(c(T,diff(test$depth)>0))

# what is first depth of increasing interval
first<-test$depth[c(1,cumsum(ind$lengths)[!ind$values])]

# what is last depth of increasing interval
last<-test$depth[cumsum(ind$lengths)[ind$values]]

# what is accumulation
last-first
[1]  1.8  4.2  3.2 -3.8
 
the similar applies to evaluation which are first and last times of events 
and/or its duration.

Regards
Petr


> 
> Thank you very much for your time.
> 
> Kara
> 
> 
> 
> 
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Look at these functions. They should give you the building blocks you  
need:

?diff   # returns a vector of length n-1
      #it would be a logical vector if wrapped in the appropriate  
functions
      #needs to be padded to line up with dataframes

?rle    # returns a list with the lengths and values of the repeats
         # will work on logical vectors

If your data is in a dataframe, snowdepth,  with column names time and  
depth, then this might  create  a difference column in that dataframe:

snowdepth$incr <- c(0, diff(snowdepth$depth))

And this might give you the be a noise marker:

snowdepth$noise <- c(FALSE, diff(snowdepth$depth) > x)

You should also look at the function dput so you can offer  
reproducible code and data with further questions.

-- 
David Winsemius

On Mar 5, 2009, at 3:46 AM, Kara Przeczek wrote:
> I am fairly new to R and I would like to do the following, but do  
> not know where to start. Any help or direction would be appreciated.
> I have a time series of snow depth measurements. I would like to  
> determine the depth of snowfall for each snowfall event. There is  
> noise in the data so I only want to add data values if the  
> subsequent depth is greater than the previous by a certain margin. I  
> am only interested in calculating snow accumulation events.
> Example data:
>
> Time	 depth	
> 1	 84.3	
> 2	 84.5	
> 3	 86	
> 4	 86.1	
> 5	 85.8	
> 6	 86.7	
> 7	 87.9	
> 8	 89.1	
> 9	 90	
> 10	 89	
> 11	 88	
> 12	 88	
> 13	 89.1	
> 14	 90	
> 15	 91.2	
> 16	 89.9	
> ...	 ...	
> I would like to create a second data frame from the data that looks  
> something like this:
>
> Event	 InitialDepth	 FinalDepth	 Accumulation	 InitialTime	 FinalTime	
> 1	 84.3	 90	 5.7	 1	 9	
> 2	 88	 91.2	 3.2	 11	 15	
> ...	 					
>
> I would like to write a program that progresses through the depth  
> column, point by point, to test if (i+1) - i > x. (where I will set  
> x to exlude the noise in the data). As long as i+1 is greater than  
> or equal to i, then the initial depth stays at the first data point  
> and the final value changes to that in i+n. Once the test is false,  
> this indicates the end of the event, the accumulation is calculated,  
> all values are saved as event X and a new event is started.
> I tried using ifelse(), but I do not know how to move through the  
> data and then save the initial and final values and time stamps in  
> another table.
>
> Thank you very much for your time.
>
> Kara
>
>
>
>
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.