Compute the variation of summary statistics for time series.
Usage
svri(
df,
bkip = NULL,
mode = "recipes",
value = "day",
st = NULL,
et = NULL,
fun = "mean",
probs = 0.5,
na.rm = TRUE,
digits = 2,
wind = FALSE,
coliws = 2,
coliwd = 3,
sn = FALSE
)Arguments
- df
dataframe of time series.
- bkip
the basic time reslution for variation, such as '1 hour'. If mode "custom" is selected, do not need to enter bkip.
- mode
for calculating cycles: "recipes", "ncycle", "custom". "recipes" means using internal setting for calculation. "ncycle" means setting number of items for per cycle. "custom" means using 1 column in dataframe as a list of grouping elements for calculation.
- value
for detail setting of mode. Possible values for "recipes" are "day", "week", "month", year". "day" equals to 24 (hours) values in 1 day. "week" equals to 7 (days) values in 1 week. "month" equals to 31 (days) values in 1 month. "year" equals to 12 (months) values in 1 year. values for "ncycle" is a number representing number of items in per cycle. values for "custom" is a number representing column index in dataframe.
- st
start time of resampling. The default value is the fisrt value of datetime column.
- et
end time of resampling. The default value is the last value of datetime column.
- fun
a function to compute the summary statistics which can be applied to all data subsets: 'sum', 'mean', 'median', 'min', 'max', 'sd' and 'quantile'.
- probs
numeric vector of probabilities with values in \([0,1]\).
- na.rm
logical value. Remove NA value or not?
- digits
numeric value, digits for result dataframe.
- wind
logical value. if TRUE, please set coliwd, coliws.
- coliws
numeric value, column index of wind speed in dataframe.
- coliwd
numeric value, column index of wind direction (degree) in dataframe.
- sn
logical value. if TRUE, the results will be presented by scientific notation (string).
Value
the variation of summary statistics Note that when the pattern USES "ncycle" or "custom", the start time determines the start time of the first element in the average variation. For example, if the first timestamp of data is "2010-05-01 12:00:00", the resolution is 1 hour, the mode is "ncycle", and the value is 24, then the result represents diurnal variation starting from 12 o'clock.
Details
If you have wind data (wind speed, and wind direction in dgree), please set 'wind' as 'TRUE', and set values for 'coliwd' and 'coliws'.
Examples
svri(met, bkip = "1 hour", mode = "recipes", value = "day", fun = 'quantile', probs=0.5,
st = "2017-05-01 00:00:00")
#> Joining, by = "temp_datetime"
#> hour of day TEM HUM WS WD
#> 1 0 17.87500 88.10000 1.700000 80.50000
#> 2 1 18.47500 91.05000 1.983333 115.66667
#> 3 2 18.09167 90.96667 1.733333 117.00000
#> 4 3 18.25000 83.11667 1.816667 112.58333
#> 5 4 18.52500 81.37500 1.766667 117.91667
#> 6 5 17.80417 88.11667 1.037500 128.70833
#> 7 6 19.70417 82.95417 1.437500 107.20833
#> 8 7 21.78750 67.86667 2.145833 128.58333
#> 9 8 22.95833 55.80833 2.687500 129.41667
#> 10 9 24.37917 51.76667 2.583333 133.25000
#> 11 10 25.02083 48.26250 2.712500 145.91667
#> 12 11 25.47500 47.15417 2.529167 139.91667
#> 13 12 25.72083 42.46250 3.241667 120.33333
#> 14 13 25.78333 37.57917 3.425000 115.29167
#> 15 14 25.90417 38.17917 3.291667 100.50000
#> 16 15 25.60833 39.17917 3.704167 84.33333
#> 17 16 24.42917 47.39583 4.183333 85.50000
#> 18 17 23.72500 52.70000 3.920000 85.30000
#> 19 18 22.21667 56.37500 2.800000 101.83333
#> 20 19 20.50833 63.16667 1.758333 96.00000
#> 21 20 19.25833 74.72500 2.116667 91.83333
#> 22 21 18.76667 77.68333 1.725000 107.75000
#> 23 22 18.52500 76.11667 1.591667 95.25000
#> 24 23 18.40000 83.22500 1.541667 131.50000