Why the CS50 Degrees Project Uses set() for Movies Instead of a List

A beginner studying the CS50 AI Degrees project may notice this line inside load_data():

people[row["id"]] = {
    "name": row["name"],
    "birth": row["birth"],
    "movies": set()
}

and naturally wonder:

Why is "movies" stored as a set() instead of a list?

After all, lists are commonly used for storing multiple values in Python.

This is actually an important design decision related to:

• data uniqueness
• graph traversal efficiency
• fast lookups
• relationship modeling

Let us understand it carefully.

---

What Does "movies": set() Mean?

This creates an empty Python set for storing movie IDs connected to a person.

Example:

people = {
    "102": {
        "name": "Tom Hanks",
        "birth": "1956",
        "movies": set()
    }
}

Initially, the actor has no movies stored.

Later, movie IDs are added using:

people[row["person_id"]]["movies"].add(row["movie_id"])

Suppose Tom Hanks acted in:

• Toy Story → movie ID 500
• Forrest Gump → movie ID 700

Then:

people["102"]["movies"]

becomes:

{"500", "700"}

---

Why Not Use a List?

A list could also store movie IDs:

["500", "700"]

So why did the project choose sets?

Because sets provide several important advantages.

---

1. Sets Automatically Prevent Duplicate Entries

Suppose the CSV data accidentally contains duplicate rows:

102,500
102,500

If movies were stored in a list:

["500", "500"]

the duplicate remains.

But with a set:

{"500"}

duplicates are automatically removed.

This is important because an actor should not be connected to the same movie multiple times.

---

2. Sets Provide Faster Membership Checking

Python sets are optimized for very fast lookups.

Example:

"500" in people["102"]["movies"]

This operation is extremely efficient with sets.

The CS50 Degrees project performs many relationship checks while running Breadth-First Search (BFS), so fast lookups matter.

---

3. Order Does Not Matter

The Degrees project only cares about:

• whether an actor appeared in a movie
• which actors are connected

It does NOT care about:

• the order of movies
• which movie came first
• sorting

Sets are ideal when:

• uniqueness matters
• order does not matter

---

How Movies Are Added to the Set

Sets use .add(), not .append().

Example:

movies_set = set()

movies_set.add("500")
movies_set.add("700")
movies_set.add("500")

Result:

{"500", "700"}

The duplicate "500" is ignored automatically.

---

List vs Set Comparison

Using a List

movies = ["500", "700", "500"]

Problems:

• duplicates allowed
• slower membership checks

---

Using a Set

movies = {"500", "700"}

Advantages:

• duplicates removed automatically
• faster lookups
• excellent for graph relationships

---

Why This Matters for BFS

The CS50 Degrees project models a graph like this:

Actor → Movie → Actor → Movie → Actor

The BFS algorithm repeatedly explores:

• which movies an actor appeared in
• which actors appeared in those movies

This requires:

• fast relationship traversal
• efficient duplicate prevention
• clean graph connections

Sets are perfect for this type of graph structure.

---

Real-World Uses of Sets

Sets are commonly used for:

• unique website visitors
• friend relationships in social networks
• unique tags or hashtags
• graph connections
• database relationship modeling

The Degrees project is essentially building a movie-actor social graph.

---

Final Insight

The line:

"movies": set()

is not accidental.

It is a deliberate design choice because:

• duplicate movie IDs should not exist
• fast membership checking is important
• ordering is irrelevant
• graph traversal benefits from set operations

This is one of the first examples where a beginner can see how data structure choice directly affects algorithm efficiency and program design.