# How can I identify a distribution?

I have data that looks like this when I plot a histogram. I have this data for several years (this is a count of values per group) and each year looks very similar.

Is there a way that I can:

1. more accurately describe this distribution (with words or with a function)

2. compare a distribution from the current year to previous ones? I'm a bit unsure what approaches are used when considering distributions but I'd like to change a process so that it looks less like the image on the left and more like the image on the right

• You have the data, right?
– Dave
Commented Oct 11, 2023 at 15:08
• Can you explain in a little more detail what you want to do? Describing, estimating and comparing distributions are three very different things. Commented Oct 11, 2023 at 15:24
• I'd say histograms are a perfectly valid approach Commented Oct 11, 2023 at 15:29
• It's usually not necessary to identify a specific distributional model, but even if it were, I wouldn't just start with a histogram. Indeed if you're using the same data as you want to perform inference on, that will carry some problems with it. Commented Oct 11, 2023 at 15:48
• I would check this page en.wikipedia.org/wiki/List_of_probability_distributions. The is the domain $[0,\infty)$, then something like Gamma or exponential distribution may be a good start. Note also that nothing is stopping you from transforming variables to create new distributions
– Cryo
Commented Oct 11, 2023 at 18:40