I have the dataset of cryptocurrency price that you can check from https://support.kraken.com/hc/en-us/articles/360047543791-Downloadable-historical-market-data-time-and-sales (I do not post a typical data here because you can the data there yourself if needed). My question is that since the intraday information of the dataset is not 'equidistant' by seconds/mintues. How can I update/calibrate it into equidistant intraday data for my purpose? Is there any quick way to do it?
Asked
Active
Viewed 13 times
-1
-
The answer depends on why you want to do that. The data is a record of events that occurred at arbitrary timepoints; if you need to quantize that to specific time periods you're either going to need to drop or merge data points; how to do that without corrupting your results is going to depend on what those results are expected to be. – Daniel Beck Aug 15 '23 at 16:44
-
@DanielBeck I understand what you mean. I am also consicous about the results that might affect the model I fit. However, the package I will use in R require the 'equidistant' timestamp. Hence I need to manipulate the data a bit without affecting too much the results, for example, I am thinking taking the near values of the nearest time record for the missing timestamp, or using the average values for a specific time interval (e.g., in every 1 min). Any suggestion? – sehne Aug 17 '23 at 08:40
-
Again, it depends on what you intend to do with the data. Nobody here can tell you which changes to the data will affect your results, without any information about what your intended results are. – Daniel Beck Aug 17 '23 at 08:52