Thursday, May 12, 2016

the long and variable project


looking at two time series on a graph ...
and i think they would look quite similar if i took the one series and lagged it

so I set up a table of dates:
lag the 2001 value to 2002
lag the 2003 value to 2004
lag the 2007 value to 2009
lag the 2011 value to 2013
lag the 2016 value to 2016

(that last one has no lag)

i am trying to have the computer figure everything out from this
to give me a graph showing the one series lagged per the table of dates.

doesn't sound conceptually difficult

but it seems to be.

i think i depend on the x-axis, the date-values, more than i realize.
i think the problem is like "the denominator problem" where you look at
for example the federal debt relative to gdp
and you see that it goes up
and you conclude that the federal debt went up
and nobody thinks about the changes in gdp.

i think it is like that.
i keep getting confused between lag-from dates and lag-to dates...
i just naturally think in terms of the y-axis values and changing those values
when really i don't want to change those values, just fit them to different dates
sometimes to impossible dates.

imp dates, i call them now. that about sums it up.

No comments: