An alternative approach to building a dashboard of your own is to make an actual web application, with a live server running Python on a backend; this will, upon request, show you a dashboard. This approach is, essentially, the exact opposite of a static dashboard in terms of pros and cons: it requires maintenance, needs to be scaled if the traffic is heavy, and could be slower. It also allows you to configure access, customize dashboards for any user or group of users, and compute the results live, even for a comparatively large dataset, without the need to share this dataset as a whole with the audience.

Of course, we could build an entire web application, controlling each and every feature (we won't do that), or use one of the specialized dashboard packages, such as uperset (essentially, a full-blown platform that requires database access) or Dash, a dashboarding tool based on the plotly Python library (very similar to Altair/Vega, but not as flexible). On the other side of the spectrum (further from static pages) are hybrid solutions, namely, panel and Voila. The latter is extremely new at the moment, so let's try building a dashboard using the panel package.