I'm having an issue graphing my model results using the jtools package in R. I am trying to plot the results of a negative binomial model which includes a logged offset, based on the jtools vignette (https://cloud.r-project.org/web/packages/jtools/vignettes/effect_plot.html) it appears that it is most appropriate to plot the partial residuals which will account for my offset but when I choose to include partial residuals my dependent variable is graphed on a log scale which makes it impossible to see the data. I've attached the code and plot for one categorical variable of the predicted and observed data with and without the partial residuals. Does anyone have suggestions for the best way to change the code so that I am still plotting accounting for the offset but able to visualize the data? I really like the figures made using this package, but if anyone has suggestions for other ways to nicely plot model predictions, that would also be helpful! Thanks!
#negative binomial glm with offset
finmod<-glm.nb(TotalAlert~Location+daynight+PropFullMoon+GroupSize+SexRatio
+PropAdult,offset=log(TotalEles),data=grouplevel2)
#plotting predictions
effect_plot(gnb6,pred=daynight,interval=TRUE,plot.points=TRUE,jitter=.2)
#plotting predictions with partial residuals
effect_plot(gnb6,pred=daynight,interval=TRUE,partial.residuals=TRUE,jitter=.2)
plot of model predictions and actual data
plot of model predictions with partial residuals and actual data
