Folks, I need to generate a time-series picture with NBER-style recession bands. This is very easy to do via the GUI, but not very much so from a script. So I put together the following function, that has a few flaws but gets the basic job done: <hansl> function void plot_with_band(list X, series band) smpl X band --contiguous scalar y0 = min(min(X)) scalar y1 = max(max(X)) scalar m = (y1+y0)/2 scalar w = (y1-y0)/1.5 series band_w = band ? w : 0 series band_c = band ? m : NA band_c[$t1] = m band_c[$t2] = m plot X options time-series with-lines options band=band_c,band_w,1 band-style=fill end plot --output=display end function ### example open AWM17 series band = (t > 1990:1) && (t<1994:4) || (t > 2000:1) && (t<2004:4) list X = CAN_YEN URX plot_with_band(X, band) </hansl> Now, I can see three possibilities: * just add this example to the User's Guide * refine the function and make it a function package * enhance the gnuplot command and the plot block with special syntax I would find number 3 preferable, if we weren't just before release, and I don't really feel like pushing release ahead for this. But OTOH, if we decide to put this off for gretl 2019c, then options 1 and 2 are not really worthwhile, are they? Ideas?

A new example in the User's Guide is always more then welcome, new gretl release or not ;))

Folks, I need to generate a time-series picture with NBER-style recession bands. This is very easy to do via the GUI, but not very much so from a script. So I put together the following function, that has a few flaws but gets the basic job done: <hansl> function void plot_with_band(list X, series band)     smpl X band --contiguous     scalar y0 = min(min(X))     scalar y1 = max(max(X))     scalar m = (y1+y0)/2     scalar w = (y1-y0)/1.5     series band_w = band ? w : 0     series band_c = band ? m : NA     band_c[$t1] = m     band_c[$t2] = m     plot X         options time-series with-lines         options band=band_c,band_w,1 band-style=fill     end plot --output=display end function ### example open AWM17 series band = (t > 1990:1) && (t<1994:4) || (t > 2000:1) && (t<2004:4) list X = CAN_YEN URX plot_with_band(X, band) </hansl> Now, I can see three possibilities: * just add this example to the User's Guide * refine the function and make it a function package * enhance the gnuplot command and the plot block with special syntax I would find number 3 preferable, if we weren't just before release, and I don't really feel like pushing release ahead for this. But OTOH, if we decide to put this off for gretl 2019c, then options 1 and 2 are not really worthwhile, are they? Ideas? -------------------------------------------------------   Riccardo (Jack) Lucchetti   Dipartimento di Scienze Economiche e Sociali (DiSES)   Università Politecnica delle Marche   (formerly known as Università di Ancona)   r.lucchetti(a)univpm.it   http://www2.econ.univpm.it/servizi/hpp/lucchetti ------------------------------------------------------- _______________________________________________ Gretl-devel mailing list -- gretl-devel(a)gretlml.univpm.it To unsubscribe send an email to gretl-devel-leave(a)gretlml.univpm.it Website: https://gretlml.univpm.it/postorius/lists/gretl-devel.gretlml.univpm.it/