Objective:
The goal of this project is to analyze Instagram engagement metrics to identify trends, detect anomalies, and derive actionable insights that can improve social media performance.
Dataset Description: The dataset consists of 119 Instagram posts with 13 columns containing various engagement metrics, such as:
Impressions (Total reach of a post)
From Explore (Views from the Explore page)
Follows (New followers gained from a post)
Likes, Comments, Shares, Saves (User interactions)
Captions and Hashtags (Text-based engagement factors)
Key Challenges & Solutions:
Outliers Detection & Correction:
Identified outliers in numerical metrics using the IQR method.
Applied capping to bring extreme values within a reasonable range to ensure fair analysis.
Data Cleaning & Preprocessing:
Checked and handled missing values.
Standardized numerical variables for better comparisons.
Exploratory Data Analysis (EDA):
Boxplots to visualize outliers before and after correction.
Histograms and KDE plots to analyze the distribution of engagement metrics.
Correlation heatmaps to understand relationships between different metrics.
Feature Engineering:
Created an Engagement Rate metric using interactions and impressions.
Identified the impact of captions and hashtags on engagement.
Predictive Analysis & Insights:
Used regression models to predict post reach based on engagement metrics.
Built visualizations to identify key trends in high-performing posts.
Expected Outcomes:
Understanding which factors contribute most to post engagement.
Recommendations for content optimization based on data-driven insights.
Improved strategy for increasing reach and interactions on Instagram posts.