Some things that could be the cause of discrepancies:
I'm using worldometers data, not Johns Hopkins
I've noticed that data sources work with different cutoffs. Meaning some end at midnight eastern, others cut off at midnight GMT, etc. Probably Worldometers has a different cutoff than JH.
Data are not exact...some sources will massage the data when they need to. For example, France data has been erratic - reporting 2 weeks of deaths on one day, having different ministries report different numbers, etc. Some orgs will try to figure out what happened...others won't.
I'm not looking at cases.
I'm looking at 7 day rolling averages of daily deaths, to smooth out the line. You may not be.
Just some ideas there.