Can I plot timeseries data in a histogram? I would like to see the distribution of a moving variable over time.
I think there might be problems because in order to see a distribution for a variable you need to take independent observations. However, timeseries data will no doubt be correlated. EG I take a 100 day timeseries for interest rate the observations would be "close" to eachother day-to-day.
Yet at the same time, the idea of a "long term" distribution for interest rate is not a ridiculous concept right? Something like the interest rate for the past 10 years I expect would have some sort of gamma distribution.
So do I just need to take observations far apart? But how far? If i take observations every month, you could argue that they're still dependent. Yet once a year would be too little.
Maybe taking a time series and mix up the data and take random samples?