CODEWITHSUNDEEP
pythonpandas
Ben Swann
Ben Swann

Reputation: 133

Calculating NPS using Pandas

I am very, very new to Python and am playing around with how I would calculate an NPS score.

The calculation is:

(count of scores 9-10/total count of scores 0-10) - (count of scores 0-6/total count of scores 0-10) for each council.

Data Frame I am using:

enter image description here

The NPS would need to be calculated for each council separately. This is my first post on here, hopefully it makes sense. If someone could point me in the right direction it would be much appreciated.

Cheers, Ben.

Upvotes: 2

Views: 2548

Answers (2)

avibrazil
avibrazil

Reputation: 330

def npsForField(df,column,fid):
    nps={}

    # first make sure our column has numeric values:
    subject = pd.DataFrame(columns=[column],data=pd.to_numeric(df[df['field_id']==fid][column]))

    # calculate all NPS components:
    nps['total']      = subject[column].count()
    nps['detractors'] = subject[subject[column]<7][column].count()
    nps['passives']   = subject[(subject[column]>6) & (subject[column]<9)][column].count()
    nps['promoters']  = subject[subject>8][column].count()

    nps['nps']        = (nps['promoters'] - nps['detractors']) / nps['total']

    return nps

Then suppose you want to calculate the NPS for a column of df called answer, but only where df[df['field_id']==fid]. Call it like this:

npsForField(df, column='answer', fid='abc123')

Sample result:

{'total': 979,
 'detractors': 313,
 'passives': 291,
 'promoters': 375,
 'nps': 0.06332992849846783}

Upvotes: 0

andersource
andersource

Reputation: 829

Assuming data is in data.csv:

import pandas as pd
from collections import defaultdict

df = pd.read_csv('data.csv')

high_nps = defaultdict(lambda: 0)
low_nps = defaultdict(lambda: 0)

high_nps.update(dict(df[df['NPS'] >= 9].groupby('CouncilName').count().reset_index()[['CouncilName', 'NPS']].values))
low_nps.update(dict(df[df['NPS'] <= 6].groupby('CouncilName').count().reset_index()[['CouncilName', 'NPS']].values))
total_nps = dict(df.groupby('CouncilName').count().reset_index()[['CouncilName', 'NPS']].values)

nps_score = {council: (high_nps[council] - low_nps[council]) / float(total_nps[council])  for council in total_nps}

print(nps_score)

Prints:

{'Council A': 0.0, 'Council B': -1.0, 'Council C': -1.0}

Upvotes: 1

Related Questions