# Plotting Binomial GLMM Probabilities in ggplot

**URL:** <https://discourse.mc-stan.org/t/plotting-binomial-glmm-probabilities-in-ggplot/30397>\
**Category:** brms\
**Tags:** techniques, bayesplot, brms\
**Created:** [February 15, 2023, 5:04pm UTC](https://discourse.mc-stan.org/t/plotting-binomial-glmm-probabilities-in-ggplot/30397 "2023-02-15T17:04:44Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Hunter24](https://avatars.discourse-cdn.com/v4/letter/h/3ab097/32.png) [@Hunter24](https://discourse.mc-stan.org/u/Hunter24)\
**Post date:** [February 15, 2023, 5:04pm UTC](https://discourse.mc-stan.org/t/plotting-binomial-glmm-probabilities-in-ggplot/30397/1 "2023-02-15T17:04:44Z")

</div>

I am trying to plot the interaction of my binomial GLMM model in ggplot.

My predictors:

Status: Male/Female  
Number of Partners: Continuous  
Status:Number of Partners

My question is using the posterior distributions, how would I calculate this?

Would it be something like this:

exp(Intercept+ Number of Partners\* values _Status(0 or 1) + Number of Partners_Status)/(1+exp(Intercept+ Number of Partners\* values _Status(0 or 1) + Number of Partners_Status))

---

<div class="post-metadata">

**Author:** ![caesoma](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mc-stan.org/caesoma/32/19767_2.png) [@caesoma](https://discourse.mc-stan.org/u/caesoma)\
**Post date:** [February 26, 2023, 1:21pm UTC](https://discourse.mc-stan.org/t/plotting-binomial-glmm-probabilities-in-ggplot/30397/2 "2023-02-26T13:21:31Z")

</div>

> [@Hunter24](#):
>
> exp(Intercept+ Number of Partners\* values _Status(0 or 1) + Number of Partners_Status)/(1+exp(Intercept+ Number of Partners\* values _Status(0 or 1) + Number of Partners_Status))

If this is the function underlying your model (or whatever else it is), you can compute its posterior distribution by computing the function for each sample in the posterior (or alternatively, for a random set of samples from the posterior). If you plot each of them as a thin line you’ll get a [spaghetti plot](https://www.nature.com/articles/s41467-020-19831-5/figures/1). A maybe more traditional way of plotting the same results is computing the functions, and at each point computing the [mean and 2.5 and 97.5 percentiles](https://www.nature.com/articles/s41467-018-03981-8/figures/1) (95% Credibility Interval, or whichever you want).