Instagram Engagement Analytics
OLS regression and caption text feature analysis identifying engagement drivers across 5,975 posts from 47 top Instagram creators.
Python (Pandas)
Overview
This project analyzes what drives engagement across 5,975 Instagram posts from 47 top creators, using OLS regression and caption text feature extraction. The goal was to isolate the impact of content format, caption characteristics, and creator identity on engagement outcomes.
Key Insights
- Album posts outperform single photos by 15.2% — the single most impactful format decision a creator can make
- Creator identity explains 45% of engagement variation, confirmed by R² jumping from 0.163 to 0.618 with fixed effects
- 🏆 Milestone Achiever Award & Finalist — Fordham Gabelli Marketing Analytics Competition 2026
Tools Used
Business Takeaway
Carousel (album) format is the highest-leverage content decision available to creators, outperforming single photos by 15.2%. The jump in R² from 0.163 to 0.618 when adding creator fixed effects reveals that brand equity and audience trust are the dominant drivers of engagement — meaning sustained audience-building matters far more than any individual post tactic.
Project Screenshots
