Customer Segmentation in Python Chapter2

DataCamp Customer Segmentation in Python
CUSTOMER SEGMENTATION IN PYTHON
Introduction to RFM
segmentation
Karolis Urbonas
Head of Data Science, Amazon
What is RFM segmentation?
Behavioral customer segmentation based on three metrics:
Recency (R)
Frequency (F)
Monetary Value (M)
Grouping RFM values
The RFM values can be grouped in several ways:
Percentiles e.g. quantiles

Pareto 80/20 cut
Custom - based on business knowledge
We are going to implement percentile-based grouping.

Short review of percentiles
Process of calculating percentiles:
1. Sort customers based on that metric

2. Break customers into a pre-defined number of groups of equal size
3. Assign a label to each group
Calculate percentiles with Python
Data with eight CustomerID and a randomly calculated Spend values.

Calculate percentiles with Python

spend_quartiles = pd.qcut(data['Spend'], q=4, labels=range(1,5))
data['Spend_Quartile'] = spend_quartiles
data.sort_values('Spend')
Assigning labels
Highest score to the best metric - best is not always highest e.g. recency
In this case, the label is inverse - the more recent the customer, the better
Assigning labels
# Create numbered labels
r_labels = list(range(4, 0, -1))
# Divide into groups based on quartiles

recency_quartiles = pd.qcut(data['Recency_Days'], q=4, labels=r_labels)
# Create new column

data['Recency_Quartile'] = recency_quartiles
# Sort recency values from lowest to highest

data.sort_values('Recency_Days')
Assigning labels
As you can see, the quartile labels are reversed, since the more recent customers
are more valuable.
Custom labels
We can define a list with string or any other values, depending on the use case.
# Create string labels

r_labels = ['Active', 'Lapsed', 'Inactive', 'Churned']
# Divide into groups based on quartiles

recency_quartiles = pd.qcut(data['Recency_Days'], q=4, labels=r_labels)
# Create new column

data['Recency_Quartile'] = recency_quartiles
# Sort values from lowest to highest

data.sort_values('Recency_Days')
Custom labels
Custom labels assigned to each quartile

Let's practice with

percentiles!
Recency, Frequency,
Monetary Value calculation
Karolis Urbonas
Definitions
Recency - days since last customer transaction
Frequency - number of transactions in the last 12 months
Monetary Value - total spend in the last 12 months
Dataset and preparations

Same online dataset like in the previous lessons
Need to do some data preparation

New TotalSum column = Quantity x UnitPrice.
Data preparation steps
We're starting with a pre-processed online DataFrame with only the latest 12
months of data:
print('Min:{}; Max:{}'.format(min(online.InvoiceDate),
max(online.InvoiceDate)))
Min:2010-12-10; Max:2011-12-09
Let's create a hypothetical snapshot_day data as if we're doing analysis recently.
snapshot_date = max(online.InvoiceDate) + datetime.timedelta(days=1)

Calculate RFM metrics

# Aggregate data on a customer level
datamart = online.groupby(['CustomerID']).agg({
'InvoiceDate': lambda x: (snapshot_date - x.max()).days,
'InvoiceNo': 'count',
'TotalSum': 'sum'})
# Rename columns for easier interpretation

datamart.rename(columns = {'InvoiceDate': 'Recency',
'InvoiceNo': 'Frequency',
'TotalSum': 'MonetaryValue'}, inplace=True)
# Check the first rows

datamart.head()
Final RFM values
Our table for RFM segmentation is completed!

Let's practice calculating

RFM values!
Building RFM segments
Karolis Urbonas
Data
Dataset we created previously
Will calculate quartile value for each column and name then R, F, M
Recency quartile
r_labels = range(4, 0, -1)
r_quartiles = pd.qcut(datamart['Recency'], 4, labels = r_labels)
datamart = datamart.assign(R = r_quartiles.values)

Frequency and Monetary quartiles

f_labels = range(1,5)
m_labels = range(1,5)
f_quartiles = pd.qcut(datamart['Frequency'], 4, labels = f_labels)

m_quartiles = pd.qcut(datamart['MonetaryValue'], 4, labels = m_labels)
datamart = datamart.assign(F = f_quartiles.values)

datamart = datamart.assign(M = m_quartiles.values)
Build RFM Segment and RFM Score

Concatenate RFM quartile values to RFM_Segment
Sum RFM quartiles values to RFM_Score

def join_rfm(x): return str(x['R']) + str(x['F']) + str(x['M'])
datamart['RFM_Segment'] = datamart.apply(join_rfm, axis=1)
datamart['RFM_Score'] = datamart[['R','F','M']].sum(axis=1)
Final result
Let's practice building RFM

segments
Analyzing RFM segments
Karolis Urbonas
Largest RFM segments

datamart.groupby('RFM_Segment').size().sort_values(ascending=False)[:10]
Filtering on RFM segments

Select bottom RFM segment "111" and view top 5 rows
datamart[datamart['RFM_Segment']=='111'][:5]
Summary metrics per RFM Score

datamart.groupby('RFM_Score').agg({
'Recency': 'mean',
'Frequency': 'mean',
'MonetaryValue': ['mean', 'count'] }).round(1)
Grouping into named segments
Use RFM score to group customers into Gold, Silver and Bronze segments.
def segment_me(df):
if df['RFM_Score'] >= 9:
return 'Gold'
elif (df['RFM_Score'] >= 5) and (df['RFM_Score'] < 9):
return 'Silver'
else:
return 'Bronze'
datamart['General_Segment'] = datamart.apply(segment_me, axis=1)
datamart.groupby('General_Segment').agg({
'Recency': 'mean',
'Frequency': 'mean',
'MonetaryValue': ['mean', 'count']
}).round(1)
New segments and their values

Practice building custom

segments

Customer Segmentation in Python Chapter2

Uploaded by

Customer Segmentation in Python Chapter2

Uploaded by

DataCamp Customer Segmentation in Python

CUSTOMER SEGMENTATION IN PYTHON

What is RFM segmentation?

Behavioral customer segmentation based on three metrics:

Grouping RFM values

The RFM values can be grouped in several ways:

Percentiles e.g. quantiles

We are going to implement percentile-based grouping.

Short review of percentiles

Process of calculating percentiles:

1. Sort customers based on that metric

Calculate percentiles with Python

Data with eight CustomerID and a randomly calculated Spend values.

Calculate percentiles with Python

# Divide into groups based on quartiles

# Create new column

# Sort recency values from lowest to highest

# Create string labels

# Divide into groups based on quartiles

# Create new column

# Sort values from lowest to highest

Custom labels assigned to each quartile

CUSTOMER SEGMENTATION IN PYTHON

Let's practice with

CUSTOMER SEGMENTATION IN PYTHON

Dataset and preparations

Need to do some data preparation

Data preparation steps

Let's create a hypothetical snapshot_day data as if we're doing analysis recently.

snapshot_date = max(online.InvoiceDate) + datetime.timedelta(days=1)

Calculate RFM metrics

# Rename columns for easier interpretation

# Check the first rows

Final RFM values

Our table for RFM segmentation is completed!

CUSTOMER SEGMENTATION IN PYTHON

Let's practice calculating

CUSTOMER SEGMENTATION IN PYTHON

Building RFM segments

r_quartiles = pd.qcut(datamart['Recency'], 4, labels = r_labels)

datamart = datamart.assign(R = r_quartiles.values)

Frequency and Monetary quartiles

f_quartiles = pd.qcut(datamart['Frequency'], 4, labels = f_labels)

datamart = datamart.assign(F = f_quartiles.values)

Build RFM Segment and RFM Score

Sum RFM quartiles values to RFM_Score

datamart['RFM_Segment'] = datamart.apply(join_rfm, axis=1)

CUSTOMER SEGMENTATION IN PYTHON

Let's practice building RFM

CUSTOMER SEGMENTATION IN PYTHON

Analyzing RFM segments

Largest RFM segments

Filtering on RFM segments

Summary metrics per RFM Score

Grouping into named segments

datamart['General_Segment'] = datamart.apply(segment_me, axis=1)

New segments and their values

CUSTOMER SEGMENTATION IN PYTHON

Practice building custom

You might also like