DataTerminology.comNEW
Also known as: Data Comprehension
A structured process of learning dataset content, quality, and context to inform modeling and decisions.
DESCRIPTION
In CRISP-DM, Data Understanding starts with initial data collection, profiling, and validation to assess suitability. Practitioners identify data quality issues, uncover patterns, and confirm assumptions. The phase bridges business goals and modeling by grounding analysis in real data constraints. Outcomes include refined problem definitions, data quality remediation, and selection of relevant variables and sources.
ALTERNATIVE DEFINITIONS
The Data Mining Group
The data understanding phase starts with initial data collection and proceeds with activities in order to get familiar with the data, to identify data quality problems, to discover first insights into the data, or to detect interesting subsets to form hypotheses for hidden information.
Defines Data Understanding explicitly as a CRISP-DM lifecycle phase with clear objectives.
Microsoft
In the data acquisition and understanding stage, the team explores, cleans, and transforms the data to ensure it is suitable for modeling.
Emphasizes practical tasks—exploration, cleaning, transformation—in an engineering lifecycle.
RELATED TERMS
CRISP-DM Data Exploration Data Profiling Data Quality Metadata Management Sampling Feature Engineering
COMPETING TERMS
KNOWLEDGE AREAS
Data Science Data Management MLOps
TARGET AUDIENCE
Data Scientist Data Analyst MLOps Engineer AI Engineer Data Steward Information Analyst
EXTERNAL SOURCES
- Wikipedia - Cross Industry Standard Process for Data Mining
- The Data Mining Group - CRISP-DM official process description
- Microsoft - TDSP lifecycle including data understanding
TOP QUESTIONS
What activities comprise Data Understanding in CRISP-DM?
Initial data collection, data description, data exploration, and data quality verification. These tasks ensure data suitability before modeling begins.
Category: Process
Why is Data Understanding critical to project success?
It aligns expectations with actual data realities, preventing wasted effort on unusable inputs. Early detection of issues reduces downstream rework.
Category: Relevancy
How does Data Understanding relate to Business Understanding?
Business Understanding defines objectives; Data Understanding tests those goals against available data, refining scope and feasibility.
Category: Concept
What tools support Data Understanding?
Data catalogs, profiling tools, notebooks, and lineage systems provide context and quality insights across sources.
Category: Tooling
What are common pitfalls in Data Understanding?
Assuming data correctness, ignoring sampling bias, and skipping documentation. Validate, profile, and record findings rigorously.
Category: Mistakes
How is Data Understanding documented?
Through data dictionaries, profiles, quality reports, and lineage diagrams, linked to business requirements and assumptions.
Category: Documentation
How long should Data Understanding take?
It varies with complexity and quality. Time-box iterations while ensuring critical risks are addressed before modeling.
Category: Management
Who owns Data Understanding deliverables?
Data scientists and analysts lead, with inputs from data stewards, engineers, and subject-matter experts, ensuring accuracy and governance alignment.
Category: Roles
What is an outcome of effective Data Understanding?
A validated feature set, defined data issues with remediation plans, and a realistic modeling path informed by data constraints.
Category: Outcome