Schema validation enforces type, range, and null constraints on incoming data before it reaches the model.
✓
Contract PASSED
All fields passed validation.
Feature
Type
Expected Range
Actual Value
Status
age
int
[18, 100]
34
PASS
income
float
[0, 1e6]
72400.00
PASS
gender
string
{M, F, NB}
M
PASS
credit_score
int
[300, 850]
720
PASS
loan_amount
float
[1000, 500000]
25000.00
PASS
employment_type
string
{FT, PT, SE, UN}
FT
PASS
has_default
bool
{true, false}
false
PASS
Data Scenario
Display Options
Failure Types
Out of range value
High null rate
Wrong data type
Invalid category
Key Insight
Data contracts fail fast at ingestion - blocking bad data before it corrupts models. Tools: Great Expectations, Pandera, dbt tests.
Data Contracts & Schema Validation - Interactive Visualization
Data contracts define the expected schema, types, ranges, and null rates for every feature entering the ML pipeline. When upstream data changes - a schema migration, a new ETL bug, or a null explosion - a contract violation is raised immediately, before the bad data corrupts the model. Tools like Great Expectations, Pandera, and dbt tests implement contracts in production. The fail-fast principle: catch data issues at ingestion, not after retraining a model on corrupted features.
Toggle between healthy and corrupted data scenarios to see which fields fail validation
See exact failure reasons: out-of-range values, high null rates, wrong data types, invalid categories
Show statistics (mean, std, null%) to understand how corruption shifts distributions
Overall contract PASS requires every field to pass - one failure blocks the pipeline
Part of the EngineersOfAI Interactive 3D - free interactive visualizations covering every major concept in machine learning and AI engineering. Hover any element for a plain-English explanation. No code required.