Marketing Data Analysis

Unlocking Advertising ROI: A Deep Dive into Social Media Campaign Performance.
This project was all about diving deep into a massive social media advertising dataset, with over 300,000 entries, to figure out what really drives a successful Return on Investment (ROI) for campaigns. My goal was to pull out actionable insights that marketers could use to make smarter decisions in the future.

What I Did and What I Found:

Getting the Data Ready: I started by meticulously cleaning and preparing the raw advertising data. This meant converting things like campaign Duration into usable Duration_Days and Acquisition_Cost into Acquisition_Cost_USD. A key step was also to break down the Target_Audience into specific Audience_Gender and Audience_Age_Group categories, which really helped in understanding who was being reached.

Exploring the Data: I conducted a thorough exploratory data analysis (EDA) to understand the patterns, find any outliers, and visualize how different campaign metrics were related to ROI. This involved creating various plots to see the distributions of age, gender, channel, and campaign type, and analyzing correlations to set the stage for building a predictive model.

Predicting Success: I then built a robust predictive model using machine learning algorithms, specifically Ridge Regression, to forecast campaign ROI. The most surprising and impactful finding was that the marketing channel itself was the strongest predictor of campaign success, even more so than traditional metrics or demographics.

Actionable Insights: Based on these findings, I developed clear, practical recommendations for businesses. These insights can help marketers optimize their social media ad spending by focusing their efforts on the channels that truly deliver the best ROI, ultimately making their campaigns much more effective.

Tools Used: I primarily used Python, leveraging libraries like Pandas for data manipulation, NumPy for numerical operations, Matplotlib and Seaborn for visualizations, and Scikit-learn for building and evaluating machine learning models (including Linear Regression, Lasso, and Random Forest Regressor).

This project really highlights my ability to take vast amounts of data, apply advanced analytical techniques, and turn those insights into tangible strategies that can drive real business growth.