HDX data check (quality and sensitivity control and data freshness checks)
All data publicly or privately shared through HDX is reviewed by a Centre for Humanitarian Data team member to ensure it follows the Terms of Service. This initial check is always performed by a member of the Centre's Data Partnerships team. A dataset will immediately be deleted from the HDX platform in case the data is found to contain personally identifiable information or other data that the Data Partnership Team member considered likely to be sensitive.
- Who to call?
- In any case, the Data Policy Officer (Jos) is notified - in case the Data Policy Officer (Jos) is not available, this notification will be forwarded to the Data Partnerships Team lead (Javier) and to the Centre Lead (Sarah).
- In case the potentially sensitive data is microdata (row-level survey data), Mike will also be notified.
- These team members, together with the Data Partnership Team member, review and document the case. After internal discussion, the contributor of the dataset will be notified and invited to jointly discuss the issue and to determine next steps.
- In case the potentially sensitive data is microdata, the following steps will be taken:
- A copy of the microdata resource is obtained by Nafi in Dakar
- Nafi applies the risk assessment tool which is part of the SDCmicro package developed by the World Bank
- Nafi shares back the risk assessment with Javier and Jos
- Javier, Jos and Nafi reach out to the contributor to inform them of the risk of reidentification of individual respondents, and explain the SDC process
- Nafi applies the SDC tool to the data concerned, and conducts another risk assessment and estimates information loss, results are shared with Javier and Jos
- Javier, Jos and Nafi determine next steps (either publication of the perturbed dataset or sharing of the original dataset through HDX connect) together with contributor
- Lessons learned are recorded in an overview of the incident, drafted by Nafi and reviewed by Jos
Further, a dataset will be made private by the Centre for Humanitarian Data team member and the contributor will be notified immediately in case the Data Partnerships Team members finds that:
● The dataset is found to contain gaps that render the quality of the dataset insufficient for use by a humanitarian audience, or;
● There are no resources (files or links) or broken links are found, or;
● The dataset contains data that is not relevant to a humanitarian audience, or;
● Data was not shared in one of the suggested formats.
After the dataset has been uploaded, the Centre for Humanitarian Data team will conduct regular checks to verify that data on the platform is being maintained. If the dataset is found to be out of date, the Centre for Humanitarian Data team will contact the contributor to suggest that the data be updated or that the dataset update frequency indication be changed. If metadata is found to be incorrect, the Centre for Humanitarian Data team will notify the contributor, after which the Centre for Humanitarian Data team might correct the metadata.