Data Quality Rules

This document describes the first public-safe data-quality checks for the Network Operations Data Lab.

The goal is to treat network documentation as structured operational data. Before this data can be used for SQL analysis or BI reporting, it should be checked for completeness, consistency and basic plausibility.

Current sample data

Current sample inputs:

• data/sample/devices.csv
• data/sample/interfaces.csv

Current device columns:

device_id, device_name, device_type, vendor, model, role, location

Current interface columns:

device_name, interface_name, interface_type, admin_status, oper_status, vlan, port_role, description

Public-safe data rule

The sample data must not contain real sensitive infrastructure details.

Do not publish:

• real serial numbers
• real MAC addresses
• real public IP addresses
• private hostnames
• Tailscale IP addresses
• customer-specific identifiers
• production configuration snippets

Use anonymized lab names and synthetic examples instead.

Device inventory checks

Recommended checks:

• device inventory exists
• device inventory is not empty
• device_id is complete
• device_id is unique
• device_name is complete
• device_name is unique
• device_type, vendor, model, role and location are documented

Interface documentation checks

Recommended checks:

• interface inventory exists
• interface inventory is not empty
• device_name is complete
• interface_name is complete
• each device_name + interface_name combination is unique
• admin_status and oper_status are documented
• port_role is documented
• description is documented where useful

Relationship checks

Recommended checks:

• every interface references a known device_name from devices.csv
• device names are consistent between both files
• no orphan interfaces exist

Reporting-oriented checks

Recommended checks:

• interfaces administratively up but operationally down
• access ports without VLAN documentation
• trunk ports without trunk documentation
• missing interface descriptions
• status and port-role summaries for BI reporting

Why this matters for BI

Power BI reports and SQL analysis depend on reliable source data.

If device identifiers, interface names, port roles or VLAN documentation are incomplete, dashboards can become misleading. These checks are intentionally simple and transparent so they can be explained in a learning, portfolio and interview context.

Current limitations

This is an early learning-lab implementation.

The current checks are based on public-safe sample CSV files. Later versions may add:

• SQLite table creation
• automated Python data-quality reports
• additional SQL validation queries
• Power BI dashboard prototypes
• documented remediation notes