SmartAd-Campaign

Objective

Test if the ads that the advertising company runs resulted in a significant lift in brand awareness. Our task here, is to design a reliable hypothesis testing algorithm for the Brand Impact Optimiser (BIO), which is a lightweight questionnaire served with every campaign.

Data

The BIO data for this project is a “Yes” and “No” response of online users to the following question

Q: Do you know the brand SmartAd? O Yes O No The users that were presented with the questionnaire above were chosen according to the following rule:

• Control: users who have been shown a dummy ad
• Exposed: users who have been shown a creative, an online interactive ad, with the SmartAd brand. The data is collected from 3-10 jul 2020 from SmartAd advertising agency.

This project is divided into four main sections:

• Creating an A/B testing framework that includes traditional, sequential, and machine learning tests
• Creating a repeatable machine learning framework
• Using MLOps best practices, conduct A/B testing with classical, sequential, and machine learning methodologies.
• Obtaining statistically valid insights in respect to the business goal

what has been implemented

• Classical Type of A/B testing
• Sequential A/B testing
• A/B testing with Machine Learning
• data versioning using DVC
• MLOps using mlfow package and Continous Integration for Machine Learning (CML).

The notebooks in this repository contains data exploration and implementation of classical p-value based algorithm,the sequential A/B testing algorithm and Machine learning based AB testing in Python.