By graphing log scale more information is revealed.  For example the first approximately 12 days the rate was greatest.  At approx. the twelfth day more effective measures were apparent. I think the breaks (15th, 21st, and 26th) are statistical and not due to changes in data collection, as was the WHO case for global data where symptom case was added instead of using only tested cases. The increase after approx. day 30 was possibly due to increasing sites.  My suppositions may of course be verified by data from other sources and the WHO situation commentaries.  

The source for this graph is WHO: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports