IanYeo's picture
Add customer LTV calculator code
5ca0c3c
from datetime import datetime
import pandas as pd
import streamlit as st
def months_between_dates(start_date, end_date):
return (end_date.year - start_date.year) * 12 + (end_date.month - start_date.month)
def calculate_lifespan(row):
if pd.notna(row["Churned"]):
return (row["Churned"] - row["Date"]).days
else:
return (datetime.now() - row["Date"]).days
def date_filtered_df(df, start_date, end_date):
return df[(df['Date'] >= start_date) & (df['Date'] <= end_date)]
def average_customer_lifespan_calculation(
df,
start_date,
end_date,
) -> float:
df.sort_values(by=['Customer', 'Date'], inplace=True)
mask = (df['Date'] >= start_date) & (df['Date'] <= end_date)
df = df.loc[mask]
df["Lifespan"] = df.apply(calculate_lifespan, axis=1)
df = df.dropna(subset=["Value"])
# Calculate average customer lifespan
return round(df["Lifespan"].mean(), 0)
def icon_select(value):
if value >= 7:
return 'πŸš€'
elif value >= 5:
return 'πŸ”₯'
elif value > 3.5:
return 'πŸ’€'
else:
return 'πŸ’€'
@st.cache_data(ttl="5m")
def get_data(file_link):
if 'dl=0' in file_link:
file_link = file_link.replace('dl=0', 'dl=1')
all_data_df = pd.read_excel(file_link)
return all_data_df
st.title('Customer LTV Calculator')
file_link = st.text_input(
'Link to data file',
)
if not file_link:
st.stop()
all_data_df = get_data(file_link)
col1, col2, col3 = st.columns(3)
with col1:
start_date = st.date_input(
'Start Date:',
value=pd.to_datetime('2022-09-01'),
max_value=pd.to_datetime(datetime.now().date()),
format='DD-MM-YYYY',
)
with col2:
end_date = st.date_input(
'End Date:',
value=pd.to_datetime(datetime.now().date()),
max_value=pd.to_datetime(datetime.now().date()),
format='DD-MM-YYYY',
)
with col3:
start_datetime = pd.to_datetime(start_date)
end_datetime = pd.to_datetime(end_date)
number_of_months = months_between_dates(start_datetime, end_datetime)
st.write(str(number_of_months), 'months')
calculated_acl = average_customer_lifespan_calculation(
all_data_df,
start_datetime,
end_datetime,
)
if start_date < end_date:
# Filter the dataframe based on the selected date range
mask = (all_data_df['Date'] >= start_datetime) & (all_data_df['Date'] <= end_datetime)
all_data_df = all_data_df.loc[mask]
else:
st.error('Error: End date must be after the start date.')
all_data_date_filtered = date_filtered_df(all_data_df, start_datetime, end_datetime)
average_order_size = all_data_date_filtered['Value'].mean()
formatted_num = "Β£{:,.2f}".format(average_order_size)
st.write('Average order size (AOS):', str(formatted_num))
purchase_frequency = all_data_date_filtered.groupby('Customer')['Date'].nunique()
average_purchase_frequency_rate = purchase_frequency.mean()/number_of_months
st.write('Average purchase frequency rate (APFR) per customer per month:', str(round(average_purchase_frequency_rate, 2)))
customer_value = average_order_size * average_purchase_frequency_rate
customer_value_formatted = "Β£{:,.2f}".format(customer_value)
st.write('Customer Value (AOS x APFR):', customer_value_formatted)
average_customer_lifespan = 12
average_customer_lifespan = st.slider(
f'Average Customer Lifespan (months) - calculated value {calculated_acl} days',
min_value=1,
max_value=50,
step=1,
value=12,
)
customer_lifetime_vale = average_customer_lifespan * customer_value
customer_lifetime_vale_formatted = "Β£{:,.2f}".format(customer_lifetime_vale)
st.write('Customer Lifetime Value (CLV):', customer_lifetime_vale_formatted)
acquisition_cost = 50
acquisition_cost = st.slider('Cost of acquisition', min_value=0, max_value=1000, step=10, value=50)
clv_cac_ratio = customer_lifetime_vale/acquisition_cost
all_data_df['year_month'] = all_data_df['Date'].dt.to_period('M')
all_data_df = all_data_df.sort_values(by='Date')
st.write(
'CLV to CAC ratio:',
"{:,.2f}".format(clv_cac_ratio),
': 1',
icon_select(clv_cac_ratio),
)