Skip to content

add text here

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
%matplotlib inline
# Import marketing.csv 
marketing = pd.read_csv('datasets/marketing.csv')

# Print the first five rows of the DataFrame
print(marketing.head())

# Print the statistics of all columns
print(marketing.describe())

# Check column data types and non-missing values
print(marketing.info())

# Check the data type of is_retained
print(marketing['is_retained'].dtype)
# Convert is_retained to a boolean
marketing['is_retained'] = marketing['is_retained'].astype('bool')
# Mapping for channels
channel_dict = {"House Ads": 1, "Instagram": 2, 
                "Facebook": 3, "Email": 4, "Push": 5}

# Map the channel to a channel code
marketing['channel_code'] = marketing['subscribing_channel'].map(channel_dict)

# Add the new column is_correct_lang
marketing['is_correct_lang'] = np.where(
    marketing['language_preferred'] == marketing['language_displayed'], 
    'Yes', 
    'No'
)
# Import marketing.csv with date columns
marketing = pd.read_csv('datasets/marketing.csv', 
                        parse_dates = ['date_served', 'date_subscribed', 'date_canceled'])

# Add a DoW column
marketing['DoW'] = marketing['date_subscribed'].dt.dayofweek
# Group by date_served and count number of unique user_id's
daily_users = marketing.groupby(['date_served'])['user_id'].nunique()

# Print head of daily_users
print(daily_users.head())

Visualizing daily marketing reach:

# Plot daily_subscribers
daily_users.plot()

# Include a title and y-axis label
plt.title('Daily users')
plt.ylabel('Number of users')

# Rotate the x-axis labels by 45 degrees
plt.xticks(rotation = 45)

# Display the plot
plt.show()

Exploratory Analysis & Summary Statistics

Calculating conversion rate

On marketing teams, conversion rate is typically the most important metric. It is one of the best ways to determine how effective a marketing team was at gaining new customers.

As a reminder, conversion rate is the percentage of the users who saw our marketing assets and subsequently became subscribers.

The formula for conversion rate is:

Number of people who convert/ Total number of people who we market to 
# Calculate the number of people we marketed to
total = marketing['user_id'].nunique()

# Calculate the number of people who subscribed
subscribers = marketing[marketing["converted"] == True]\
                            ['user_id'].nunique()

# Calculate the conversion rate
conversion_rate = subscribers/total
print(round(conversion_rate*100, 2), "%")

Is this a good conversion rate? This will depend heavily on the business. There is no particular number that all marketing teams are trying to reach. It helps to look at historical data to determine whether a conversion rate is what you would expect.

# Calculate the number of subscribers
total_subscribers = marketing[marketing.converted == True].user_id.nunique()

# Calculate the number of people who remained subscribed
retained = marketing[marketing.is_retained == True].user_id.nunique()

# Calculate the retention rate
retention_rate = retained / total_subscribers
print(round(retention_rate*100, 2), "%")
‌
‌
‌