What is an example of a data set you have worked with, will work with, or are familiar with which could be modeled with GLMs or mixed-effects models?
Very briefly(!), what is the nature of the question and the structure of the data?
How would you first plot the data?
What are some simple models you could start with?
What might a more complete model look like? What intercepts or slopes might you allow to vary across groups (be treated as "random effects")?
What are some issues you might expect to encounter? Temporal or spatial autocorrelation? Heteroscedasticity? Group-level predictors? Overdispersion? How will you check for these?