Data Science Fundamentals
Data exploration, cleaning, and preparation.
Data Cleaning and Preprocessing
Handling noise, inconsistencies, and formatting issues – garbage in, garbage out is the first law of ML.
Data Splitting and Sampling
Train/validation/test splits, stratification, and handling class imbalance – the foundation of honest evaluation.
Data Types and Structures
Numerical, categorical, ordinal, text, time series – understanding your data’s nature determines every downstream decision.
Encoding Categorical Variables
One-hot, label, target, and embedding-based encoding – translating categories into numbers without introducing false relationships.
Exploratory Data Analysis
Visualizing distributions, correlations, and anomalies before modeling – the most undervalued step in the ML pipeline.
Feature Scaling and Normalization
Standardization, min-max scaling, and robust scaling – ensuring features contribute equally regardless of their original units.
Handling Missing Data
Deletion, imputation, and model-based approaches – the strategy depends on why data is missing, not just how much.