Exploring how actors are connected through shared film collaborations.
Interactive Application
Explore shared films and discover the shortest collaboration path between actors.
Project Overview
Movies are created through large networks of actors who repeatedly collaborate across different productions. However, a standard movie table does not clearly show how individual actors are connected or how collaboration patterns extend across films.
This project uses graph analysis to make those relationships visible. In the network, each actor is represented as a node, and an edge connects two actors when they appeared in the same movie.
Problem Statement
Movie datasets commonly organize information by title, release date, popularity, and cast list. While this format is useful for looking up individual movies, it does not make it easy to explore relationships between actors.
How might I transform movie and cast data into an interactive network that helps users explore actor relationships, shared movies, collaboration paths, and influential performers?
Project Goal
Dataset
The project uses two CSV files from the TMDB 5000 Movie Dataset: movie metadata and movie credits with cast information. The files include movie identifiers, titles, release dates, popularity values, vote averages, cast members, and crew information.
Network Structure
Each node represents an actor.
An edge connects two actors who appeared in the same movie. Each edge stores shared movie titles and a weight representing how many movies the actors made together.
My Role
Process
I identified actor collaboration as a relationship that could be better understood through a graph rather than a traditional table.
I loaded and merged the TMDB movie and credit files, parsed cast records, removed duplicate actor names within each film, and handled empty or invalid values.
I represented actors as nodes and their shared movie appearances as weighted edges.
I created functions for actor search, shared-movie lookup, rankings, actor details, network summaries, and shortest-path analysis.
I transformed the command-line functionality into a Streamlit application with sidebar navigation and separate views for each feature.
I wrote automated tests to verify graph construction, search, actor details, path finding, rankings, invalid input handling, CSV loading, and network summaries.
I completed the Python network model, Streamlit application, test suite, and project documentation.
Object-Oriented Design
Stores movie ID, title, release date, vote average, popularity, and cast members. The class also derives the release year from the release date.
Stores actor name, movie titles, unique collaborators, movie count, and collaborator count.
Manages the NetworkX graph, actor and movie objects, data loading, search, rankings, shared movies, collaboration paths, and network statistics.
Core Application Features
Users can enter part of an actor's name and receive matching results. Search normalizes capitalization and extra spacing.
Users can view an actor's movie count, unique collaborator count, film titles, and collaborator names.
NetworkX determines how two actors are indirectly connected through other performers and shared films.
The program ranks actors by the number of unique performers with whom they collaborated.
Users can enter two actors and identify the movies in which they appeared together.
The application reports total actors, collaboration edges, movies, and connected components.
Most Connected Actors
A careful interpretation is that high connectivity may reflect a combination of career longevity, number of films, ensemble casts, and the dataset's coverage, not necessarily artistic influence by itself.
Testing
I wrote automated tests to verify graph construction, search, actor details, path finding, rankings, invalid input handling, CSV loading, and network summaries.
Reflection
This project helped me connect programming fundamentals with analytical thinking by turning relationship data into a useful exploration tool.
Behind the Project
Curious how the network comes together? Visit the GitHub repository to explore the source code, data workflow, automated tests, and documentation behind the interactive experience.
