Netflix Movies & TV Shows Analysis

Goal

Analyze the Netflix titles dataset to answer:

• Movies vs TV shows split
• Top countries producing content
• Content additions per year
• Popular genres/categories

Analysis pipeline

diagram Netflix catalog pipeline mermaid

  flowchart LR
  A["Raw titles
(CSV)"] --> B["Clean
(parse dates, split country/genre)"]
  B --> C["Explode
(one row per country/genre)"]
  C --> D["Visualize
(type split, top countries, yearly adds)"]
  D --> E["Conclude
(catalog trends)"]

From raw titles to content trends by type, country, and year.

Common columns (Netflix titles dataset)

• type (Movie/TV Show)
• title, director, cast
• country, date_added, release_year
• rating, duration, listed_in

Step 1: Load and parse dates

Load dataset

import pandas as pd
 
df = pd.read_csv("data/netflix_titles.csv")
 
if "date_added" in df.columns:
    df["date_added"] = pd.to_datetime(df["date_added"], errors="coerce")
 
print(df.shape)
print(df.head())

Step 2: Clean multi-valued country/genre

The dataset often stores multiple values separated by commas.

Split and explode

# Countries
if "country" in df.columns:
    countries = (
        df.dropna(subset=["country"])
          .assign(country=df["country"].str.split(","))
          .explode("country")
    )
    countries["country"] = countries["country"].str.strip()
 
# Genres
if "listed_in" in df.columns:
    genres = (
        df.dropna(subset=["listed_in"])
          .assign(genre=df["listed_in"].str.split(","))
          .explode("genre")
    )
    genres["genre"] = genres["genre"].str.strip()

From the book

McKinney parses the Bitly "a" (user agent) field by splitting on whitespace and keeping just the first token, because the raw string packs too much into one cell. country and listed_in have the same problem here — one cell holds several values joined by commas. str.split(",") + explode() turns each into its own row so value_counts() can count them correctly, one country/genre at a time.

Step 3: Movies vs TV shows

Type counts

import seaborn as sns
import matplotlib.pyplot as plt
 
plt.figure(figsize=(6, 4))
sns.countplot(data=df, x="type")
plt.title("Movies vs TV Shows")
plt.tight_layout()
plt.show()

Step 4: Top countries

Top countries

import seaborn as sns
import matplotlib.pyplot as plt
 
if "country" in df.columns:
    top_c = countries["country"].value_counts().head(10).reset_index()
    top_c.columns = ["country", "count"]
 
    plt.figure(figsize=(10, 4))
    sns.barplot(data=top_c, x="count", y="country")
    plt.title("Top 10 countries by number of titles")
    plt.tight_layout()
    plt.show()

Step 5: Additions over time

Titles added by year

import matplotlib.pyplot as plt
 
if "date_added" in df.columns:
    yearly = df.dropna(subset=["date_added"]).groupby(df["date_added"].dt.year).size()
 
    plt.figure(figsize=(8, 4))
    plt.plot(yearly.index, yearly.values, marker="o")
    plt.title("Titles added by year")
    plt.xlabel("Year")
    plt.ylabel("Count")
    plt.tight_layout()
    plt.show()

Visualize it

sketch Movies vs TV Shows split p5.js Pause Restart

A donut chart makes the catalog mix obvious at a glance.

Deliverable

Summarize 5–10 insights:

• Which type dominates?
• Which countries dominate?
• Any clear growth periods?
• Which genres show up most?

🧪 Try It Yourself

Exercise 1 – Split a multi-value column

pch.exerciseTag pch.exerciseTitle

Exercise 2 – Explode into one row per value

pch.exerciseTag pch.exerciseTitle

Exercise 3 – Titles added per year

pch.exerciseTag pch.exerciseTitle