Recovering Short-Interval Volume from Rolling 24-Hour Totals

Summary

The document describes a data problem common in cryptocurrency markets: exchanges may report a rolling 24-hour traded volume at each observation, while analysis often needs volume over shorter intervals such as a minute. It asks whether to compare the rolling total at the start and end of a bar after adjusting each for the total from 24 hours earlier.

The text provides the proposed calculation as a question, but no answer, test, or worked sample. Conceptually, differencing rolling-window totals can recover interval activity only when timestamps and window boundaries are aligned and the observations needed to remove the outgoing 24-hour slice are available. Sparse or irregular sampling, delayed updates, and exchange-specific volume definitions can complicate reconstruction. The document therefore frames the issue but does not establish a standard method or validate the proposed procedure.

Key ideas

• The input series contains rolling 24-hour volume totals rather than interval volume.
• The author proposes differencing rolling totals while accounting for activity leaving the window.
• Recovering short-interval volume requires aligned timestamps and information about the outgoing window segment.
• The document offers no answer or empirical validation of the proposed method.

Tags

# Transform 24hr cumulative volume to sampled periods


# Transform 24hr cumulative volume to sampled periods












I have a Python Dataframe with cryptocurrency data that has three columns: time, 24hr volume and price.

The time is the time at which the data was received from the exchange, price is the last price of the currency at that time and volume is the total traded volume from that point in time back until 24hrs ago.

I would like to be able to find what the volume is for sampled period such as: total volume of a minute

Here is what I think needs to be done:

Lets say I want to sample my data to 1 minute volume bars.

Take the volume at the close of the bar and subtract it by the volume from 24hrs ago. Then take the volume at the open and subtract it by the volume from 24 hours ago. then I find the difference of the two numbers I have just obtained and that would be my volume traded over the minute?

Is this the proper way to do it? unfortunately a lot of cryptocurrency exchanges only give you what the 24hr cumulative volume was, but volume is more useful over sample periods such as 1,5 or 10 minutes.

Is there a standard way of doing this?

Sample data:

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.