Every friendship leaves a data trail.
From Friends to Features transforms real social connections into structured insights using thoughtful data analysis.
From Friends to Features is a social data analysis project built on raw data collected from friends, their mutual connections, and the pages they engage with. The goal of the project is to understand how social proximity and shared interests can be used to generate meaningful recommendations.
By cleaning, processing, and analyzing real-world social data, this project demonstrates how connections and preferences can be translated into features that reveal potential friendships and interest-based recommendations.
- Source: Self-collected data from friends and their mutuals
- Data Includes:
- User connections
- Mutual friends
- Pages liked by users
- Nature: Raw, unstructured, and user-generated
The dataset required careful preprocessing to make it suitable for analysis.
- Python
- JSON
-
Data Collection
- Gathered raw social data from friends, their mutual connections, and liked pages
-
Data Cleaning & Preprocessing
- Handled missing and inconsistent entries
- Removed duplicates and noise
- Standardized user and page representations
-
Data Transformation
- Structured social relationships into analyzable formats
- Created features representing mutual connections and shared interests
-
Data Analysis
- Identified similarity patterns between users
- Analyzed overlap in liked pages
- Studied the impact of mutual friends on social proximity
- Suggested pages users may be interested in based on shared preferences
- Identified potential friend recommendations using mutual connections
- Revealed how overlapping interests strengthen social links
- Demonstrated how small social networks still contain rich patterns
- Real-world data collection
- Data cleaning & preprocessing
- Feature engineering from social data
- Exploratory data analysis
- Recommendation-style reasoning
- Analytical problem-solving