MarketPulse: Building an AI Marketing Analyst for Small Businesses

This submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle aims to create an AI Marketing Analyst called MarketPulse. The application is designed to help small business owners, freelance marketers, and marketing teams make sense of campaign performance data and make better-informed decisions. MarketPulse aggregates several tools into one platform, including campaign analytics, marketing insights, data quality checks, and a budget planner.

The main features of MarketPulse include interactive charts for campaign and channel performance, trend identification and areas needing attention, data issue detection, and budget allocation exploration. It also offers a FutureCast forecasting feature and an AI Marketing Assistant that answers questions about campaign metrics and provides explanations and recommendations based on the available data.

The AI Marketing Assistant uses the open-weight openai/gpt-oss-20b model for its explanations and suggestions, accessed through Groq's hosted inference API. The application combines deterministic calculations with language-model assistance, ensuring that the AI's explanations are accurate and relevant to the presented marketing context. Throughout development, the author encountered the importance of combining AI assistance with reliable calculations and validation, as well as clear limits on the model's capabilities.

MarketPulse is still under development and validation, but the author has tested it using marketing campaign data and applied their experience running Ghifty's Bowl, their fruit-salad business. The application is built using Python and Streamlit, with additional technologies such as Pandas, NumPy, and Plotly. The open-weight model chosen for MarketPulse allows developers to experiment with AI-powered products while maintaining data analysis and engineering control over the AI's output.