Python Quantified Self Dashboard - Should I use Jupyter notebook?

I’m currently trying to start building my Python application for Quantified self Dashboards & Correlational analysis.

Would you recommend using Jupyter notebook? Would Jupyter Notebook help me build this application?

The purpose of the application I’m building is to aggregate all my tracking data (From .CSV files, API’s, JSON files etc.) at runtime for dashboards and analytics. At first it will probably only be correlational analysis, but later maybe also machine learning.

Your question is a few days old but I figured it might still be nice to share my experience as I’ve been using python and Jupyter Notebook for this since months.
I’d advise to code with jupyter and to run the scripts „stand-alone“ on startup with command line via python interpreter. That’s more stable and more resource efficient. Works for me without problems. The Keyword is „cronjobs“.

More ressource efficient and stable than what? And in what way? :slight_smile:
Do you then have UI?

I need to stress right ahead that I’m no IT-Guy but I assume you are planning to automate everything and won’t need to recode the script every day.

  1. in terms of memory efficiency: When using jupyter you’d always need to have a browser running as well. But you can simply „convert“ the notebook to a .py script and run that script without wasting resources on a browser via a bash script for example. (Checkout nbconvert with jupyter via Google)

  2. visualization: Jupiter notebook is a coding environment and not an actual application to run although you could use it as such. I mean you could use jupyter for this purpose but that’s pretty much like using excel for a presentation instead of PowerPoint. To visualize everything in a neat manor I’d suggest to look into the Tkinter or dash library for a GUI.

One of the advantages of using Jupyter notebooks as a coding environment is that there are active projects going for sharing notebooks for analyzing personal data acquired via wearables APIs, like and the Personal Data Notebooks collection on Open Humans: