That's because this error is caused by yt doing something incorrect when interfacing with NumPy internally. Whenever yt raises a mysterious exception like this from something its doing internally it's almost certainly due to a bug in yt or yt failing to convert the internal error into a more understandable user-facing error.
First, try updating yt, the newest version is 3.5.1 and it's possible that your issue has already been fixed.
If that doesn't work, can you make a full runnable example that triggers this issue? If you can, use one of the public test datasets on yt-project.org/data
but if its somehow related to properties of the data you are working with, you can share a dataset using the "yt upload" subcommand:
$ yt upload some_file.tar.gz
That will print out a URL you can share once its done uploading.
Once you have that you should open a bug report on yt's github including the full example that triggers the issue: