Once information requirements and relevant data standards have been identified, organisations need to decide on the quality of data required to ensure the data is fit-for-purpose. Data quality dimensions defined by the Data Management Association (DAMA) include:
- Accuracy
- Accessibility
- Completeness
- Consistency
- Integrity
- Reasonability
- Timeliness
- Uniqueness/de-duplication
- Validity
Without a sufficient level of confidence in the data, an accurate view of the infrastructure and operations is incomplete, which may lead to poor decision-making.
As a general recommendation, preference data quality over data quantity. It is better to have a well-structured data set that has been verified and validated, and is reliable, than a large volume of poorly organised, unreliable data.
Data Quality Requirements
Data quality issues
Data quality issues caused by device breakdowns or device calibration can generate incorrect or inaccurate data which can lead to incorrect decision making. If this poses unacceptable business or customer risk, agencies should use their contract to define data governance requirements and required mitigations that minimise the likelihood of these risks. These can include service level agreements with service providers for fault identification, remediation and re-calibration of devices at regular intervals; acceptable standards for data quality; uptime and availability requirements; and consistency of data over time.
Liability arising from data quality
Agencies need to be transparent about any potential quality issues in licence or sharing agreements if the data will be made available to others as open data or as shared data. It may be important to flag in any contracts or sharing agreements that data may be incomplete, intermittently available or otherwise unreliable if there are connectivity or outage issues impacting your network. This will help protect against any liability claims.
Contracts, data licences and data sharing agreements must make clear that the NSW government is not responsible for any liability issues that may arise from data quality issues or reliance by users. NSW government organisations must be transparent about any quality issues and have high quality, routine, and well-governed processes in place to ensure the timeliness and accuracy of infrastructure data. This will mitigate against the likelihood of any impactful data quality issues occurring.
To guard against any liability issues that may arise with the use of a third-party product derived from NSW government data, agencies should seek legal advice on appropriate wording and include a disclaimer in any licence agreements. Disclaimers will not completely eliminate risk, but a combined metadata statement, licensing agreement and disclaimer is a suitable method for risk mitigation.
Last updated 12 Nov 2020