Exploratory Data Analysis (EDA)

EDA is the process of getting to know your dataset before building any models. You use descriptive statistics and summary functions to uncover patterns, anomalies, and distributions. EDA answers questions like: "Are there outliers?" "Which columns are most correlated?" "Is the data balanced across groups?"

• df.describe() — summary statistics (count, mean, min, max, quartiles) for numeric columns.
• df.info() — column names, data types, and non-null counts. Essential for spotting missing data.
• df['column'].value_counts() — frequency distribution of categorical values.
• Always visualize distributions early — a histogram or boxplot reveals outliers and skew that summary stats alone can hide.

Practical Skill: In Jupyter Notebooks, run EDA commands cell by cell. Use Pandas for tabular summaries and Matplotlib or Seaborn for quick plots. The goal is to form hypotheses, not confirm them.

# Exploratory Data Analysis with Pandas          # Install: pip install pandas           import pandas as pd          import numpy as np           # Create a mock public health dataset          data = {              "clinic": ["A", "B", "C", "D", "E", "F"],              "patients": [340, 510, 480, 180, 620, 280],              "staff": [5, 8, 6, 2, 10, 4],              "region": ["Central", "Central", "Equatoria", "Equatoria", "Central", "Upper Nile"],              "vaccination_rate": [0.72, 0.88, 0.65, 0.91, 0.95, 0.54]          }           df = pd.DataFrame(data)           # 1. Structure overview          print("=== Dataset Info ===")          print(df.info())           print("\n=== Summary Statistics ===")          print(df.describe())           print("\n=== Per-Region Distribution ===")          print(df["region"].value_counts())           print("\n=== Average Patients by Region ===")          print(df.groupby("region")["patients"].mean().round(0))           # Key finding          print("\nInsight: Central region has more staff and higher patient volume.")          print("Upper Nile has the lowest vaccination rate — may need investigation.")