search

gwastro / pycbc

Core package to analyze gravitational-wave data, find signals, and study their parameters. This package was used in the first direct detection of gravitational waves (GW150914), and is used in the ongoing analysis of LIGO/Virgo data.
http://pycbc.org
GNU General Public License v3.0
315 stars 354 forks source link

Chain / acf Plots for Ensembles Should Default to Mean Over Walkers #1807

Open farr opened 7 years ago

farr commented 7 years ago

It is clearer to interpret and easier to see trends when plots like chain plots (ACF plots, etc) show the trend of the mean position of the ensemble versus time rather than an individual track for each walker. (For the ACF, the ACF of the mean walker position is equivalent to the mean of the ACFs for each walker, and will have much reduced noise if the samples are really decorrelated.)

This is true for all ensemble samples (kombine, emcee, emcee_pt).

It is, of course, reasonable to add an option like --show-all-walkers to plot individual walker trajectories when the user wants to check for, e.g. stuck walkers, but it shouldn't be the default.

cdcapano commented 7 years ago

PR #1832 addressed the first half of this (for plot_acf). We still need the ability to take the mean over walkers for plot_samples. @farr @cmbiwer if you'd like to take a stab at that sometime the next couple of weeks, feel free. Otherwise I'll try to get it after I get back from the PAX conference in two weeks.

ahnitz commented 3 years ago

@cdcapano Is this still an issue?