Have the following simplified dataframe:
Date Name Score V H M
2018-01-01 A 5 V1 H4 M6
2018-01-01 B 3 V5 H2 M1
2018-01-01 C 4 V7 H6 M6
2018-01-01 A 4 V11 H9 M3
2018-01-01 C 2 V4 H2 M18
2018-01-02 A 4 V9 H1 M9
2018-01-02 B 1 V15 H4 M10
2018-01-02 A 3 V10 H10 M14
2018-01-03 C 5 V5 H21 M34
2018-01-04 A 3 V8 H9 M6
2018-01-04 A 4 V4 H15 M9
2018-01-04 C 2 V2 H4 M14
2018-01-04 B 5 V1 H1 M2
Looking at the above dataframe as the raw set, I've further indexed it by date and resampled at a Monthly level. What I'd eventually like to do is create individual timeseries for unique values in the columns (Name, V, H, M) with respect to the score (accomplished by grouping). While I've downsampled, I expect irregular timeseries sizes across these columns' grouped unique values and plan to interpolate to handle that.
The goal is to create and extract multiple time series into a new dataframe and to explore their correlation maps. For example, I would have individual time series for V1, V2, ... , Vn, H1, H2, ... , Hn, M1, M2, ... ,Mn, and so on.
I'm not sure if this should be all captured in one new dataframe or multiple dataframes based on the groupings. Here's an example of what the timeseries output should look like:
Date Score
V1 2018-01-01 4.5
2018-02-01 4.1
2018-03-01 4.3
2018-04-01 4.2
2018-05-01 4.4
Date Score
V2 2018-01-01 4.5
2018-02-01 4.1
2018-03-01 4.3
2018-04-01 4.2
2018-05-01 4.4
Date Score
V3 2018-01-01 4.5
2018-02-01 4.1
2018-03-01 4.3
2018-04-01 4.2
2018-05-01 4.4
I need help in implementing an efficient way to do this and to know if I'm on the right track. The dataframe above is a simplified version of a larger dataset.
Appreciate any help and guidance.