Hi there,

I have plotted my brms results like this:

```
myPlot=stanplot(myFit,pars=c("p1","p2","p3"),prob=0.95,prob_outer=0.95)+
scale_y_discrete(labels=c("label1","label2","label3"))
```

Is it possible to change the colour of just the first bar? I’ve tried

```
+ scale_color_bayesplot(values=c("red","black","black"))
```

but that does nothing

I’d be very grateful for any advice

Jacquie

Unfortunately, the `color`

aesthetic is not bound to anything in `stanplot`

or the underlying `mcmc_intervals`

call, so you cannot do it directly. The quickest way to do it manually is probably using the `summary`

function (below a quick hack in `tidyverse`

style for fixed parameters only):

```
summary(fit)$fixed %>%
as.data.frame() %>%
rownames_to_column("parameter") %>%
ggplot(aes(color = parameter, y = parameter)) +
geom_segment(aes(x = `l-95% CI`, xend = `u-95% CI`, yend = parameter)) +
geom_point(aes(x = Estimate) )
```

If you need both 50% and 95% intervals you either need to cleverly combine two `summary`

calls or directly use and process the result of `posterior_samples()`

. `posterior_samples()`

would also allow you to easily work with all parameters at once (without referencing throuhg `$fixed`

or `$random`

) and access the actual random effects (`$random`

only gives you the `sd`

parameters).

Hope that helps.

thank you so much, that’s very helpful! I was not quite sure where I could find out how it works and whether I’m just missing something as I’m still getting used to R…

In case this might be useful for others: Martin’s suggest above works well. There is also a beautiful package (@strengejacke) that does it and that I didn’t see before: https://strengejacke.wordpress.com/2017/10/23/one-function-to-rule-them-all-visualization-of-regression-models-in-rstats-w-sjplot/