Skip to main content
OnePocketLab
Sign in
onepocketML
Beginner

Data & Preprocessing

Real data is messy. Learn to clean it, shape it and trust it.

The unglamorous work that makes every model possible, handling missing values, encoding categories, scaling features and splitting data honestly before it ever reaches an algorithm.

8 of 8 lessons available

RAW RECORDS BECOME A USABLE TABLE???AS COLLECTEDREADY TO TRAINcleanencodescaleGAPS, WRONG TYPES AND WILD OUTLIERS FIRSTMOST OF A PROJECT HAPPENS HERE
Gaps, wrong types and wild outliers become a table a model can read. This is where most of a project's hours actually go.

Lessons in this path

  1. Understanding Your Dataset: Rows, Columns and Data Types
  2. Handling Missing Data
  3. Encoding Categorical Data: One-Hot and Label Encoding
  4. Feature Scaling: Normalization and Standardization
  5. Feature Engineering: Creating Better Inputs
  6. Detecting and Handling Outliers
  7. Train/Test Split and Cross-Validation in Practice
  8. Working with Imbalanced Datasets
Data & Preprocessing | OnePocketML