Score
0
Level
1
Streak
0
Accuracy
0%
Time
00:00
Lives
❤❤❤
Level progress0 / 5 puzzles
Choose an encoding method
Select a card, then solve the mapping puzzle.
One-hot
Creates a separate indicator column for every category.
NominalOrdinal
Maps meaningful ranks to ordered numeric values.
OrderedLabel
Assigns one integer label to each category.
CompactBinary
Converts category indices into compact binary columns.
High cardinality
Select a method and start the puzzle.
Drag category tiles into matching encoded slots, or enter mappings manually.
Edit values before encoding. Missing values may be written as blank, NA, or null.
Nominal versus ordinal
Nominal categories have no natural rank. Ordinal categories have a meaningful order, such as low, medium, and high.
False ordering risk
Label encoding can make models treat arbitrary numbers as distances. One-hot encoding avoids that assumption.
Train-test consistency
Fit category mappings on training data only. Reuse the same mapping for validation and test data.
High cardinality
Binary encoding uses fewer columns than one-hot encoding. It is useful when a feature has many unique values.
Method comparison
Analytics dashboard
Plotly charts update after every puzzle.
Results and diagnostics
Complete a puzzle to see diagnostic feedback.
Achievements
Keyboard and accessibility
Use Tab to focus method cards. Press Enter to select. Canvas tiles can also be solved using mapping controls.