Manganite has been created to empower master and doctoral students in econ and management to turn research notebooks into interactive dashboards. The students use Python for data analysis, math programming, and basic machine learning. Instead of pouring their results into slides, posters, and papers to communicate findings, they can now use Manganite to create interactive dashboards without having to leave their familiar Jupyter environment.
Manganite is built on top of the powerful Panel framework [1], but it does not require the user to convert their existing code into callbacks. Instead, Manganite parses the Jupyter notebook and builds a dependency tree between its cells. Scalar variables and Pandas DataFrames are transparently wrapped in Param classes, which lets them watch their values for changes and propagate these changes downstream.