New Github libraries
This blog post introduces four new GitHub libraries, each a result of personal research and practical needs, aimed at providing robust solutions in their respective fields.
1) ADVECT: A Leap in Marketing Campaign Analysis
ADVECT stands for Advertising Effectiveness & Conditional Treatment Analyzer, a sophisticated tool designed to estimate the Average Treatment Effect on the Treated (ATT) employing the Conditional Independence Assumption (CIA). Tailored for marketing campaign data analysis, ADVECT enhances strategic planning and assessment, enabling precise targeting and resource allocation. A real-world application saw a national wine merchant utilise ADVECT to optimise an advertising campaign, resulting in increased memberships and sales through targeted household analysis.
Usage: ADVECT is accessible via a Jupyter notebook, guiding users through power analysis, data preprocessing, and ATT estimation, ensuring campaigns target the most responsive audiences effectively.
2) MarketMix Shapley Analyser (MMSA): Refining Marketing Performance Analysis
The MarketMix Shapley Analyser (MMSA) applies Shapley Value Regression to deconstruct marketing performance at a detailed level, surpassing traditional Marketing Mix Modeling (MMM). This tool allows for the identification of each channel partner's unique contribution, providing granular insights into marketing strategies.
How to Use: MMSA comes with a step-by-step guide, from data formatting using a provided template to running analysis through a Jupyter notebook, culminating in a comprehensive evaluation of marketing partners' performance.
3) Monkey Portfolios Revisited: Testing Financial Theories
Inspired by Burton Malkiel's provocative suggestion in "A Random Walk Down Wall Street," the Monkey Portfolios Revisited repository empirically tests the theory that randomly selected portfolios can compete with expertly curated ones. This exploration includes portfolio construction and analysis, providing a modern take on the efficient market hypothesis.
Getting Started: The repository features a Jupyter notebook for simulating and evaluating "monkey" portfolios, inviting users to examine the impact of random stock selection on portfolio performance.
This section of the repository highlights the application of convex optimization techniques for portfolio construction, offering methodologies for objectives like weight minimization and entropy maximization. This approach allows for innovative portfolio strategies that balance risk and return efficiently.
The newly released GitHub libraries represent significant advancements in applying data science across marketing, finance, and data analysis. These tools are designed to enhance efficiency, provide deeper insights, and explore theoretical concepts with practical applications. From improving marketing campaign effectiveness to challenging established financial theories, these libraries open new avenues for research and strategy development in various domains.
Explore these innovative solutions and contribute to the evolution of data-driven approaches across industries.
Visit the GitHub repositories here: https://github.com/credenceanddecision
