For new digital, data collected externally to statistical offices, the principle of moving computation out towards the sources (instead of moving raw data into the statistical office) is among the pillars of Trusted Smart Statistics. The design of data processing methods for computing statistics remains under the responsability of statistical offices. Such methods are more complex to design than those for traditional data, and they will partly executed outside the statistical office, at the data source. In this scenario the processing methods must be described in terms of open-source software code in addition to more abstract human-readable public manuals.