Exploratory Data Analysis (EDA) on Titanic

Goal

Perform EDA on the Titanic dataset and produce:

Data quality findings (missing values, types)
A handful of clear plots
Insights about survival patterns

Analysis pipeline

Every EDA project follows the same rhythm: load the raw file, fix the messy parts, ask simple questions of the clean data, draw a picture, then write down what you saw.

diagram Titanic EDA pipeline mermaid

  flowchart LR
  A["Raw CSV
(Titanic passengers)"] --> B["Clean
(fill Embarked, flag Cabin)"]
  B --> C["Explore
(value_counts, groupby)"]
  C --> D["Visualize
(survival by sex/class/fare)"]
  D --> E["Conclude
(who survived, and why)"]

From raw CSV to survival insights, one stage at a time.

Dataset

Common sources:

Kaggle: Titanic - Machine Learning from Disaster

Typical columns:

Survived, Pclass, Sex, Age, SibSp, Parch, Fare, Embarked

Step 1: Load data

Load Titanic CSV

import pandas as pd
 
df = pd.read_csv("data/titanic.csv")
print(df.shape)
print(df.head())

Step 2: Schema and missingness

Info + missing

print(df.info())
 
missing = (df.isna().mean() * 100).sort_values(ascending=False)
print(missing)

Focus on missing in:

Age
Cabin
Embarked

Step 3: Clean minimal issues

Handle missing Embarked (small)

Embarked fill

if "Embarked" in df.columns:
    df["Embarked"] = df["Embarked"].fillna(df["Embarked"].mode().iloc[0])

Keep Cabin as “has cabin” flag

Cabin flag

if "Cabin" in df.columns:
    df["has_cabin"] = df["Cabin"].notna()

Step 4: Univariate plots

Survival distribution

Survival count

import seaborn as sns
import matplotlib.pyplot as plt
 
plt.figure(figsize=(6, 4))
sns.countplot(data=df, x="Survived")
plt.title("Survival counts")
plt.tight_layout()
plt.show()

Age distribution

import seaborn as sns
import matplotlib.pyplot as plt
 
if "Age" in df.columns:
    plt.figure(figsize=(7, 4))
    sns.histplot(df["Age"].dropna(), bins=30, kde=True)
    plt.title("Age distribution")
    plt.tight_layout()
    plt.show()

Step 5: Survival by category

Survival by sex

import seaborn as sns
import matplotlib.pyplot as plt
 
plt.figure(figsize=(7, 4))
sns.barplot(data=df, x="Sex", y="Survived")
plt.title("Survival rate by sex")
plt.tight_layout()
plt.show()

Survival by passenger class

Survival by class

import seaborn as sns
import matplotlib.pyplot as plt
 
plt.figure(figsize=(7, 4))
sns.barplot(data=df, x="Pclass", y="Survived")
plt.title("Survival rate by class")
plt.tight_layout()
plt.show()

Step 6: Numeric relationships

Fare vs survival (boxplot)

Fare vs survival

import seaborn as sns
import matplotlib.pyplot as plt
 
plt.figure(figsize=(7, 4))
sns.boxplot(data=df, x="Survived", y="Fare")
plt.title("Fare vs survival")
plt.tight_layout()
plt.show()

Visualize it

A pivot_table collapses “survival by sex” and “survival by class” into a single matrix — the same trick McKinney uses to get mean movie ratings by gender in one line instead of several separate groupby calls.

Survival matrix

matrix = df.pivot_table("Survived", index="Pclass", columns="Sex", aggfunc="mean")
print(matrix)

sketch Survival rate by class and sex p5.js

Taller amber bars mean a higher survival rate for that class/sex group.

Step 7: Write insights (example)

Write 5–10 bullet insights such as:

Survival rate is higher for females.
Higher class passengers survived more.
Passengers who paid higher fare tended to survive more.
Missingness is high in Cabin; treat as a feature (“has_cabin”).

Deliverable

Save a cleaned dataset version:

Save output

df.to_csv("output/titanic_cleaned.csv", index=False)