clean up test_*.py sample data to match sample public data set

[MichaelTiemannOSC]

It's all well and good to have small, stable examples of companies, ISINs, production, emission, and intensity data. But these data should tie back to some realistic initial sample data. e.g., we should not use the CISCO ISIN to describe an Electricity Utility supposedly located in France, nor an ISIN from France for a utility based in Germany, etc., as is observed in test_base_providers.py and others.