Need Pandas Correlation for multiple column with respect to dates as index

Question

m2 = m.groupby('month_year')[
    ['units', 'revenue', 'transactions', 'Sessions', 'Bing Ads',
     'Criteo','Facebook', 'Google Ads', 'Google Shopping']
].corr()

This is what i am doing but getting nan idk why?

See output results for your reference here

Added raw data

Asked By: Syed Farhan Ahmed

||

Source

Answer 1

From the raw data you added, I don’t see shared dates between observations. Here’s an example:

data = [
    ["day 1", 88, 23, 20],
    ["day 2",-5, 78, 14],
    ["day 3",8, 35, 14],
    ["day 4",-3, 33, 155],
    ["day 5",88, 28, 80],
]


df = pd.DataFrame(data, columns=["id","a", "b", "c"])
df.groupby("id")[["a", "b","c"]].corr()

Here there are 5 days, none in common, as your raw data shows. This will return a df of NAs:

          a   b   c
id                 
day 1 a NaN NaN NaN
      b NaN NaN NaN
      c NaN NaN NaN
day 2 a NaN NaN NaN
      b NaN NaN NaN
      c NaN NaN NaN
day 3 a NaN NaN NaN
      b NaN NaN NaN
      c NaN NaN NaN
day 4 a NaN NaN NaN
      b NaN NaN NaN
      c NaN NaN NaN
day 5 a NaN NaN NaN
      b NaN NaN NaN
      c NaN NaN NaN

If you have dates in common instead:

data = [
    ["day 1", 88, 23, 20],
    ["day 1",-5, 78, 14],
    ["day 1",8, 35, 14],
    ["day 2",-3, 33, 155],
    ["day 2",88, 28, 80],
]


df = pd.DataFrame(data, columns=["id","a", "b", "c"])
df.groupby("id")[["a", "b","c"]].corr()

Will result in:

                a         b         c
id                                   
day 1 a  1.000000 -1.000000  1.000000
      b -1.000000  1.000000 -1.000000
      c  1.000000 -1.000000  1.000000
day 2 a  1.000000 -0.924148 -0.147259
      b -0.924148  1.000000 -0.241780
      c -0.147259 -0.241780  1.000000

From the images posted this is my conclusion.

Answered By: Chris

Need Pandas Correlation for multiple column with respect to dates as index

Question:

Answers: