LatAm vs Nigeria: is headcount comparable?

The Copenhagen series resets completely in 2003, with no bridge data, and the notes say "new methodology adopted." Without overlap data, there's no way to establish whether the new series is at the same level or offset by a constant. Should we flag this as a comparability break, or do we have bridge-year data somewhere else? See /u/plg_simon_proudfoot/p/plot-0004.

Why is the weekend slice so much smaller? Is something wrong with the day-of-week parsing?