@ruud - yes - you’re correct. I updated the single Overview dashboard file in v1.05 if you want to grab that. Or just edit the file you downloaded and switch my silly “air_st” to “obs_air” in both locations. Save it - and restart the Grafana docker image. I’ll take a look at the Wind details in the Today so far and rerun the log file into it.
It’s easy enough to change the metric details as everything in the logs are actually already metric. For example, in the query:
avg_over_time({app="weatherflow"} |= "obs_air" | json obs_Air_Temperature="obs[0][7]" | unwrap obs_Air_Temperature | __error__="" [$__interval]) * 9/5 + 32
Just remove the " * 9/5 + 32 " at the end of the line. Then change the Units from Fahrenheit to Celsius. The issue comes from - because I built an “All In One” - it also means the dashboards provision themselves every time Grafana starts up and you can’t save your changes unless you make a new dashboard. (Which you totally can do that). But the database for Grafana then keeps track of all of those changes that resides in the Docker container. If you pull a new update or recreate your container, you lose your changes. That gets into more data retention, storing data into mapped Docker volumes, then permissions, etc. Which takes something intended to be “just start it up” to something customized to each person’s familiarity with running each of the components. At that point, learning how to deploy and manage makes more sense and just loading in your edited dashboards makes a lot more sense.
As far as the wind direction arrow - I’m not a fan of it either. But there’s really no good usable compass rose widget. Swapping out arrows is the best I can do until some of the other community Grafana visualization plugins get updated. There’s another project that has a nicer-looking compass but it’s not compatible with the data that I’m using.
Lastly - I’ve been working with building out a REST API collector, as well as a Forecast collector, and all of these multiple device/API types, are going to have some breaking changes in the data collector. That kind of stinks for historical data collection. Every day I learn/think about something new and it kind of makes a mess of the backend. Each backend stores and makes visualization different. 90% of them are super each to just make work - the other 10% makes it next to impossible to keep things the same. I need to settle on one and just use it without bouncing all over the place 
But I guess that’s how these projects work. Making one project do everything increases the complexity a lot more than I was expecting and my time is pretty limited at the moment. If only I had a set of business and functional requirements to guide my tinkering.
Certainly appreciate community enthusiasm. It’s all a work in progress!! Thanks!