plot inconsistencies - Githubissues

quantifish / rlsd

The LSD R package

Other

1 stars 0 forks source link

plot inconsistencies #29

Open merrillrudd opened 1 year ago

merrillrudd commented 1 year ago

[x] text size and axis label angles -- sex ratio residuals, sex ratio fits, cpue residuals, cpue fits
[ ] LF violin plots - shading based on the sum of weights during each year or size -- shading based on size is constant, is this a function of LF standardisation or a bug?!
[ ] Colour palette -- when colored by season, AW should be cooler color (see LF violin plots)
[ ] On all residual plots should add dashed horizontal lines at -2, 2
[ ] Some plot labels x-axis is Year while for others it is Fishing year. Should make consistent
[ ] Likewise for size - sometime we just use Size (mm), other times we use Midpoint of size-class (mm), and in the comparison figures it is Length bin

quantifish commented 1 year ago

Season legend is surplus to requirements - make git issue for next yearAlso - the previous plots violins are shaded based on the sum of the weights during each year, this should be done in this plot too (but not this year)Also - not sure who changed the colour palette for this one, but it should be changed back to be the same as all the other plots or vice versa. AW should be the colder colour too!

Just a note to myself to address this before plenary -- current code is trying to calculate weight by size for shading, but they are all equal. maybe something to do with the new LF standardisation?

the color was previously by source, but now source is no longer there. changed back to the color palette that matches the LF fits. AW is the warmer color in the other plots too sooo that can be an issue for later if you feel very strongly next year :-P

quantifish commented 1 year ago

need to make a note somewhere that these plots need to be tidied up if we are going to use them in ppts etc. A git issue

quantifish commented 1 year ago

selectivitiy plot max should be 1

quantifish commented 1 year ago

U plot should read "Exploitation rate (U)" on y axis - i.e., not caps "Rate"

quantifish commented 1 year ago

In the future we should group MCMC figures into blocks of related pars. For example:

Likelihoods and priors Productivity pars like growth and M Selectivity and vuln pars