Cohort analyser
This script is made to make customer cohorts of varying lengths and sample their order behavior. It will determine the percent of orders placed in windows of varying size from customer signup. As well as the percent of those orders which are first time orders.
To run this script 2 csv files are expected as arguments.
CSV Format for customers
customer_id,created_date
CSV Format for orders
id,order_number,user_id,created
Simple execution of this script would be as follows:
./analyze.py --customers customers.csv --orders orders.csv
Sample Execution of the script with increased cohorts plus html output
./analyze.py --customers customers.csv --orders orders.csv --cohorts 10 --html-output output.htmlSample help output from script:
$ ./analyze.py -h
usage: analyze.py [-h] --customers CUSTOMERS --orders ORDERS
[--cohort-length COHORT_LENGTH] [--cohorts COHORTS]
[--bucket-length BUCKET_LENGTH] [--html-output HTML_OUTPUT]
Perform customer order cohort analysis.
optional arguments:
-h, --help show this help message and exit
--customers CUSTOMERS
customer csv file path
--orders ORDERS orders csv file path
--cohort-length COHORT_LENGTH
cohort lengths in days (default 7 days)
--cohorts COHORTS number of corhorts to analyze (default 8)
--bucket-length BUCKET_LENGTH
bucket length in days(default 6 days)
--html-output HTML_OUTPUT
html file to output results table to)