"
- Build end-to-end digital systems to reduce errors and protect the integrity of the data across the federal enterprise.
- Build once, use often.
- Tap into the "golden sources" of data. Don’t rely on derivative databases or data derived from other data sources. Go directly to the transactional systems that do the business on a day-to-day basis.
- Release data in a machine-readable format and encourage third-party applications.
- Employ common data standards. Think about what would have happened if railroads across the country had used different standards in terms of railroad track gauges.
- Use simple, upfront data validations.
- Release data as close to real time as possible.
- Engineer systems to reduce burdens.
- Protect privacy and security. This is critical, especially in the age of Facebook and Twitter. You can create a mosaic effect without really thinking about it. It’s one thing to release data when it comes to health care on a state level, and other thing to release it on a zip-code level.
- Provide equal access to data and incorporate user feedback on an ongoing basis."
No comments:
Post a Comment