[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
"Nick Cox" <n.j.cox@durham.ac.uk> |

To |
<statalist@hsphsun2.harvard.edu> |

Subject |
st: RE: overlaid graphics |

Date |
Mon, 16 Feb 2004 11:29:35 -0000 |

It seems to me you don't want to overlay two time series as such. You just want two scales, one a multiple of the other. Here's another way to do that. Let's read in some time series data . use http://www.stata-press.com/data/r7/grunfeld.dta, clear If I pick a time series . twoway connect mvalue year if company == 1 and I also find out its initial value of 3078.5, I can then use a utility -mylabels- from SSC which is designed for these problems: * you want to show labels on one scale, but * the graph is working with values on another scale. In this case, values range from just below to almost twice the initial value. I decide on a base of 100 and index values 80 100 120 140 160 180, whereas the graph is working with values which are those multiplied by 3078.5/100. That is, label I want scale used by -graph- ------------------------------------ 80 80 * 3078.5 / 100 100 100 * 3078.5 / 100 etc. I put the desired labelling information in a local macro: . mylabels 80(20)180 , myscale(@ * 3078.5 / 100) local(labels) -mylabels- echoes the macro, so I can see when I'm on the right track: 2462.8 "80" 3078.5 "100" 3694.2 "120" 4309.9 "140" 4925.6 "160" 5541.3 "180" A suitable graph is then something like . twoway connect mvalue year if company ==1 , yaxis(1 2) yla(3000(500)6000, ang(h) axis(1)) yla(`labels', ang(h) axis(2)) yli(3078.5, axis(2)) ytitle(initial value = 100, axis(2)) If you just want fractions, then the call is clearly . mylabels 0.8(0.2)1.8 , myscale(@ * 3078.5) local(labels) Nick n.j.cox@durham.ac.uk John D. Levendis > > I've got a question regarding how to overlay two time series. > > Let say I've got a Price Series (Pr) and from it, I generate > a Price Index (PI) > that just divides Pr by it's base year value: > . gen PI = Pr/Pr[1] > > Appropriately scaled these two series, when overlaid should > match up perfectly. > What I would like to do is graph both these series, and if I > look at the units > on the left hand side of the graph, I get the Price values, > and if I look at > the units on the right hand side, I get the Index values. > > I thought something like: > . twoway (connected Pr year) (connected PI year, yaxis(2)) > > But they don't line up exactly. How can I force them to? > > This problem holds even in the simpler case where: > . gen Pr2 = Pr + 100 > . twoway (connected Pr year) (connected Pr2 year, yaxis(2)) * * For searches and help try: * http://www.stata.com/support/faqs/res/findit.html * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**st: Output of bs-command: References for use of normal/percentile/bias-corrected confidence interval** - Next by Date:
**RE: st: How to estimate var-cov matrix from data** - Previous by thread:
**st: Output of bs-command: References for use of normal/percentile/bias-corrected confidence interval** - Next by thread:
**RE: st: How to estimate var-cov matrix from data** - Index(es):

© Copyright 1996–2014 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |