6. Maintenance and updates
This step can be implemented as follows
The data administrator and person promoting data sharing
- Update the shared data observing the specified update rate or as needed, for example when an error is detected
- Also update the metadata of the published data
The person promoting data sharing together with the communications unit
- Inform data users of changes as necessary
This part describes how the metadata, or descriptive data, for the opened dataset can be maintained.
No official recommendations exist for maintaining the metadata for datasets to be opened.
Organisations that have already opened their data have striven to specify who is responsible for the metadata, maintenance and monitoring of the opened dataset. It is advisable to designate the party responsible for the dataset and include this information in its metadata. This responsibility may, for example, be assigned to a unit or a role. The responsible party carries out tasks related to the metadata, maintenance and monitoring of the dataset.
It is advisable to update the datasets or the interface observing the update rate specified in the metadata. Where an interface is used, the data may be automatically updated on a continuous basis, whereas a dataset may be updated annually, for example. If possible, the data should be updated automatically.
It is important for the data administrator to inform data users or potential users through the interface of any changes to be made to the interface, either using different communication channels or directly, if the API key provides information on the interface users.
If data is shared through an API, any changes to the content will be visible to the users in the API implementation. In practice, a new version of the interface is created. When versioning the interface, attention should be paid to publishing both backward compatible and incompatible modifications and corrections. At the same time, the organisation should note that it might need to provide support for several interface versions simultaneously. Any changes to the interface must be described in the metadata.
The national API principles should be followed when maintaining interfaces. It is important to create a process for maintaining interfaces and have indicators for monitoring interface use. Read more about the API principles.
If the dataset is updated manually, its metadata must be checked and updated at the same time with the data. For example, if statistics are updated annually, the metadata must also be updated every year. If, for some reason, the dataset needs to be modified before the date of the next update, the metadata should be checked and, if necessary, updated at the same time.
Some data portals provide the data administrator with the option of receiving an automated e-mail notification when the date specified for the next update in the metadata of the opened dataset approaches. This helps the data administrator ensure that the dataset and its metadata are up to date.
Practices of Helsinki Region Infoshare
The metadata of the data must be updated at least as frequently as the data itself. The metadata must indicate when manually updated data has been updated and when the next update can be expected. The metadata for automatically updated data should also be reviewed from time to time to check that it matches the data content.
The organisation should provide information about data updates in the same way as it communicates about the opening of new datasets, at least in the case of major updates. If possible, it is advisable to communicate about such procedures as interface version updates in advance.
Practices of the Finnish Environment Institute
The Finnish Environment Institute uses a tool for describing the metadata (metadata editor) and has a separate metadata service for end users.
The metadata is always updated together with the dataset. The organisation checks at least once a year that the metadata for continuously updated datasets is up to date. The responsibility for producing the metadata and keeping it up to date rests with the persons responsible for the data and the system administrators of information systems.
Both the metadata editor and the metadata service offer open interfaces for harvesting metadata. This means that any modified data are automatically updated in the Paikkatietohakemisto service maintained by the National Land Survey of Finland (metadata compliant with the INSPIRE Directive), the opendata.fi service maintained by the Digital and Population Data Services Agency (metadata for Finnish Environment Institute’s open datasets) and CSC's Etsin.fi service.
Updating opened data
This part describes how needs to update the open data, which may be related to data quality or errors in the dataset, can be managed.
No official recommendations exist for managing the update needs of datasets to be opened, but the Information Management Board has issued a set of recommendations for applying certain information security statutes (in Finnish) (Ministry of Finance publications 2021:65), according to which information risk management is a continuous activity, and the information management entity should describe the objectives, responsibilities and key methods related to it.
Updating open datasets and responding to possible problems
A new version of the data is available
Always keep the data up to date.
Even if a new version of the data is published, the old version does not necessarily need to be removed. Historic data makes, for example, yearly comparisons possible.
Improving dataset quality
The quality of published data should be constantly improved.
There may well be room for improvement in the quality of the opened dataset, which is why the data administrator should encourage data users to report any problems, shortcomings or weaknesses in the data quality. Users can submit their proposals for improving the data quality.
The data administrator should respond as quickly as possible to any shortcomings or problems identified and take the necessary corrective action.
Users can be invited to contact directly the party specified in the metadata, such as the contact person, customer service or registry of the organisation that opened the data.
The data administrator can, for example, improve the completeness of the data by providing it in several different file formats.
Correcting an error detected in a dataset and communicating about it
The data administrator should prepare a plan in advance for the eventuality that an error is detected in the opened dataset, or that the dataset contains personal data or other sensitive information, despite data protection and information security checks. It is advisable to specify the responsible persons and measures to be taken in such situations in the plan. It is also important for the organisation to plan and agree on how data protection and information security issues relevant to the opened datasets are monitored and developed.
According to the Information Management Board’s set of recommendations for the application of certain information security regulations (in Finnish) (Ministry of Finance publications 2021:65), the management is responsible for the organisation of, and allocation of resources to, information risk management. In addition, the information management entity maintains datasets comprising the risk assessment results and risk management plans and regularly assesses if this data is partly or fully secret or classified.
In problem situations involving data protection and information security, the data administrator should react immediately. The data administrator must remove the dataset (from the administrator’s website and/or the data portal), contact the organisation's data protection officer and information security experts, and launch information activities following the pre-prepared plan.
It is also important for the data administrator to react quickly to other errors detected in the dataset. The data administrator should rapidly assess how and when the organisation can take the necessary corrective actions and inform users about this following the communication plan. The data administrator can either remove the dataset completely until further notice or update the metadata with information about the identified error and announce when the error will be corrected.
The preparation of a crisis plan should be integrated into the organisation's crisis communication plan, if any. To prepare this plan, VAHTI cooperation network’s publication Management of information security incidents (in Finnish) VAHTI 8/2017 can be used.
Response to problem situations at Statistical Finland
The primary objective of correcting errors in statistical data shared on online services or in printed publications is to guarantee that correct data is used and ensure that incorrect data is corrected as effectively as possible. Transparent correction procedures also help to maintain user confidence in the producer of the statistics.
If significant errors are detected in a statistical publication or the online service, the publication guideline for the Official Statistics of Finland (OSF) is followed in the procedure for correcting them.
Information about a material error should be spread at least as widely and visibly as about the original data. However, if the incorrect data has attracted a great deal of publicity, it may be necessary to publish a separate notice about the correction with wider distribution and higher visibility than what was given to the original data.
The information about the error must remain permanently visible, except for in database tables. Significant errors and the dates of their correction should be added to the page describing changes in the statistics.
Publishing and notices about data publication
Significant errors occurring in connection with the publication of a dataset are corrected by publishing a notice about the correction, which is distributed as widely and visibly as the incorrect data. In the online service, the error should be corrected as quickly and visibly as possible. A separate reference to the correction procedure should be added, for example to the page describing changes in statistics.