Why is the data from the same source (KTTD Portland/Troutdale Airport) reporting different numbers from separate providers?
I get that there may be a polling time difference, but shouldn’t they line up after a reload of those entities?
I’m collecting data from Ambient Weather network (KTTD) and the National Weather Service (NWS). Here are some snapshots showing data, and some historical charts:
My goal is to track, as accurately as possible, sailing conditions and the inaccuracies between the many sources astounds me. Clicking on the same location, using Windy.com is even farther from these numbers. The difference between 3knts and 14knts is huge.
Finding reliable source in the Azores can be a challenge, but I’m planning a sail there and the numbers are way off between providers.
I’m curious if there are methods (manual polling) that may increase the consistancy such that I can pic a source and provider.
I don’t think this has anything to do with HA or the integrations, but rather variations from providers. For fun, I just looked up PDX on both of the providers you used and the numbers were just as different. Interestingly, Ambient Weather was claiming the last update was from an hour ago.
I would recommend just using a single provider for all locations for maximum consistency.
What I’m trying to understand is if I should throw out the standard refresh request rates and use an automation to force a query, and if that would even make a difference.
I get that different providers/integrations may be sampling at their own rates. What I would like to understand is if I setup an automation to sample at a designated rate, if it would even make a difference. I suspect there is some caching going on at the provider, and if setting up an automation would get more realtime data.
Also know that when you reload your GUI, then you only re-read the data in the HA state machine.
You do not get new data from the providers, which is handled by the integrations and often the update interval is locked to a specific value.
