Am 23.11.2021 um 23:16 schrieb Hanku Lee
<hanku.lee(a)gmail.com>:
Most fitness enthusiasts have favorite apps to track activities and
share with friends. Often, advanced features need a paid membership.
Still, it is limited to fixed features and I get distracted or let
down by certain features, which I don't need. Feature mismatch
persists.
I'm plenty annoyed with many fitness tracker manufacturers who force consent to store
all data with the manufacturer after purchase in order to use the device at all. Others
allow only very basic use without consent to data storage. In fact, others offer only very
sparse and inflexible data evaluation options. This even applies to Apple, which are quite
elaborate after all. And above all, they offer very limited export and exchange options,
unless you want to resort to proprietary and manufacturer-owned data repositories.
There is no technical reason to store the data on your own hardware. You "only"
have to take it in hand.
Fedora Server with its reliable service and in combination with other open source tools
(e.g. R for graphical and statistical analysis), and all that power on an SBC would be a
wonderful, affordable appliance.
I think this would be an excellent and especially innovative use case.
I have 7+ years of fitness data in Strava. The main sport is
cycling.
I want to keep the data in my control and design a custom dashboard
using the Fedora server and R Pi. It's not huge data, so I'll stay
away from the cloud. This is proof of concept if I want to expand this
further in the future.
I'm most attracted to "data in my control“. :-)
[Objectives]
Build data pipeline with fitness API (start with Strava).
Set up a database server on R Pi.
Write SQL scripts for data ingest and metrics.
It’s a question of details, but instead of SQL scripting you might use an
object-relational mapping tool. In java it would be Hibernate, for Python there are
equivalent libraries, I think.
...
(Optional) Expand data source to fitness sensors (cycling computer and
Heart rate monitor) for richer metrics.
I suspect that would be the "hard part'"
Are you sure that this is an open source project? If it already says "pricing"
in the main menu …
Again, in Java I know several open source projects (e.g. based on Netbeans infrastructure
that includes powerful backend capabilities). But certainly there are a lot of truly open
source projects in Python as well.
A very interesting project.