r - Can I make a time series work with date objects rather than integers? -
r - Can I make a time series work with date objects rather than integers? -
i have time series info i'm trying analyse in r. provided csv excel, subsequently read data.frame all
. let's has 2 columns: all$date
, all$people
, representing count of people on particular date. frequency hence daily.
being excel, dates integers representing number of days since 1900-01-01. read info people = ts(all$people, start=c(all$date[1], 1), frequency=365);
gives silly start value of 40000 because info starts in 2006. start
parameter doesn't take date object, according ?ts
, can't utilize as.date()
:
ts - ... start: time of first observation. either single number or vector of 2 integers, specify natural time unit , (1-based) number of samples time unit. see examples utilize of sec form.
i of course of study set start=1
, it's bit painful figure out season we're in when plot tells me interesting things happening around day 2100. (to clear, setting frequency=365
tell me year we're in, isn't useful more precise dates). there useful way of expressing date in ts
in human-readable form don't have maintain calling as.date()
understand when interesting features happening?
r date time-series
Comments
Post a Comment