mc_forecast dosen't work because historical observations depends on AGWeatherNet which doesn't work about farmvibes-ai HOT 5 CLOSED

Hello, please also bear in mind that we need historical data from the weather station for which you want to make more accurate forecast. Relying on OpenWeather or other data source that have low spatial resolution could lead to a degradation in performance.

from farmvibes-ai.

Thanks for reaching out for the issue identified.
The deepmc showing the example to execute the notebook using AgWeatherNet station (Palouse) data. Based on station name the path of source is derived. Please make sure to have the data available at below path prior to execution.

./data/{STATION_NAME}/training.csv

Attaching the sample data downloaded from AGWeatherNet.

palouse_agweathernet.csv

For the missing package "unfoldN", we will verify the missing packages.

from farmvibes-ai.

Thank you ionnes , indeed it's our goal to use deepmc to drive our weather station predictions more over we are trying to train the system to the errors between weather station actual's & predictions versus weather services actual's & predictions in the hopes to deliver a superior weather service eventually , I think that the system offers an opportunity for correlating spatial & temporal errors in weather services after enough training

from farmvibes-ai.

How does AgWeather register a 'No value' value? Thanks in advance,

from farmvibes-ai.

Hi, @aialves. Are you seeing any issue with the workflow itself due to no values?
The AgWeather API documentation has additional information on how they handle these values.

from farmvibes-ai.