Time series comparison

Is there a way to compare two distinct time series? The aim is to see also visually if there are concomitant or close events. The point is that the two data series don’t have the same time base. I’ll explain: - I have two data series extracted from different DB on the same period of time (last year and half);

  • Series 1 has its own time base;
  • series 2 has its own time base as the data are acquired independently of the data relating to the first time series.
    The line plot has the limit of being able to use only one time base at a time, otherwise I could have displayed the two time series on the same graph.

Hi @Guglielmo,

sorry but I could not understand why you are not able to plot both on the same chart.
Can you make an example or upload some data? :slight_smile:

Thank you,
RB

1 Like

Lello,
Node like Line Plot allow to use only one column as “x” axis so you can compare data series with same granularity and same date-time basis. My problem is a bit more complicate I want to compare date with different granularity and different bases of date-time. The two time series is referred to the last year and half but they have different date-time bases (and different granularity).

Hopefully the problem now is more clear.
I cannot upload data as they are company data … I should manipulate data in order to trasform them in “anonymous content” but if the problem description is clear I wolud like to avoid manipulation work.

Many thanks for any suggestion.

Hi Guglielmo,

if data have the same granularity, I’d aggregate one dataset so it matches the second one. Doing the opposite (disaggregating one dataset to match the granularity of the second one) does not seem a great solution to me.

As regards the example dataset, keep in mind that uploading the whole dataset is not interesting; we just need the dates/periods so we can better understand how to proceed and a fake value so we can see some results on plots. I don’t want to peek inside your data, it’s just that I can’t visualise the solution without knowing what we are dealing with :slight_smile:

Have a nice weekend,
RB

Hi Lello,
data sets have different granularity and in addition different Date&time bases. I’ll try to prepare a workflow and clean the data because I’m afraid that my explanation is not completely clear.

I will back as soon as possible.

Regards

Below a simple workflow and relative data sets
demo_group.knar (137.8 KB)

Actually I found a solution. I joined the two series accordingly to the Data-Time columns and after I merged the two Data-time column in order to have just one Data-Time column and generate time series with same length … there are some missing value to manage but it’s works.

If you have different solution you are welcome. Sorry for my questions but I’m still new of this complex tools.

Thank you

Regards

Guglielmo

2 Likes

Sorry but yeaterday I was sick and could not access the forum. Glad you found a solution :slight_smile:
It sounds as a good one!
Have a nice week,
RB

1 Like

You might have also tried to use datetime shift node to move both to the same timeframe but the different granularity would still be to cope with
br