This post provides a good explanation why and how the data collection process should be part of the implementation of an open data project rather than just a once-executed-and-then-forgotten step. Ideally the data collection building scripts plus various processing steps, e.g., conversion to various data formats, dataset merging, or parsing of meta data are made a open source project themselves. See this open-source project for an example.
I am interesting in other such projects for various reasons:
- Learn how to interpret various types of open data collections
- Best practices for open data "build" systems