Open
Description
JetBrains' Datalore is another collaborative data science platform, like Deepnote. From @densnow's preliminary investigations, it seems the pros and cons are:
- Datalore has better code/Markdown support, including linting and grammar checking, and making improvement suggestions.
- Datalore only has two sharing levels (view or edit), while Deepnote has intermediate commenting level.
- One must be logged in Datalore to view the shared notebooks, while DeepNote has a 'everyone with link' sharing mode that requires no account.
- Deepnote notebooks are in a database, whereas Datalore notebooks are in the file system and so we can run tools (nbqa, nbconvert, black, etc.) on them, which makes type checking more accurate, for example.
- Like Deepnote, Datalore allows publishing the notebook, as a 'report'. The report can be viewed publicly and can be commented on by Datalore users. Comments on the report are visible in the original notebook, but not vice versa. This allows authors to add personal comments (eg to-do notes) not visible to others.
It thus seems worthwhile to investigate Datalore further as a possible option and to see whether a similar pre-built project could be set up for copying. If so, the documentation would need updating.