estradafelipe / voucher_recommendation

Voucher Recommendation

Objective

This project attempts to show the interplay between different frameworks in the field of data. The tools that are used in this project are:

• Docker - containerisation and standardisation of processes
• Postgres - Database to store data

This project uses Python, SQL and BASH.

Data Cleaning

The dataset was filtered and cleaned according to the following criteria:

• Only entries holding the Peru country code were included
• Entries with total_orders of 0 or null were discarded.
• Entries with last_order_ts that took place before first_order_ts were discarded.
• Entries with null voucher_amount were discarded.

Requirements

In order to run this project, Docker needs to be installed and running. Furthermore, ensure that the ports used in the docker-compose are not blocked by other processes you might already have running.

Steps to execute

The repository's root directory contains an executable quickstart.sh . While being in the root directory execute the following to initialise the project: ./quickstart.sh

The executable runs docker-compose, activate de venv, execute etl job and the API

API enpoints

A REST API is connected to the database and provides access to information via the following endpoints:

• http://localhost:8000/voucher - endpoint that accepts requests and returns a JSON object
• http://localhost:8000/all - endpoint that displays all recommended vouchers for all segments