With MeteoShrooms, you can keep track of the perfect conditions for mushroom hunting. We take Open Government Data (OGD) from the official Swiss Meteorological Service MeteoSwiss to calculate key metrics.
|
ℹ️
|
Contributors Welcome! See Contributing Guidelines for more info… |
While hunting mushrooms happens in nature, a lot of the planning can be carried out beforehand.
Where are possible areas to find king boletes?
Should you leave the house right now or are you already too late?
And will your current small basket be large enough or is it time to invest in something bigger?
Meteorological data, time series of past findings as well as other datasets can give us valuable insights.
The main goal is to develop a model to predict mushrooms occurences using OGD data. I suggest starting with meteo data by MeteoSwiss, soil properties (see for example data.geo.admin.ch or opendata.swiss) and using open mushroom observation data from GBIF.
Currently, the MeteoShrooms project is focussed on Switzerland, as the Swiss meteorological agency MeteoSwiss has started to provide access to their data as Open Government Data (OGD).
A main aim is to have sound results calculated efficiently. Therefore, Python Polars is used for calculations, wherever possible. Its LazyFrames optimize the workflow. Data quality is validated through Pandera. Compressed Parquet-Files save storage space and up-/download bandwith.
Currently, the repository hosts all parts of the modelling process: 1. Data Preparation, 2. Model Calculations and 3. Data Presentation:
-
Data Preparation: Download data from OGD sources and bring them into form needed for next step. Ensure data quality,
-
Model Calculations: Perform the calculations to generate predictions from observed data,
-
Data Presentation: Show results on map and/or dashboard. Include documentation of sources and methodology.
flowchart TD
ground[(Ground Properties)] --> prep
meteo[(Meteo Data)] --> prep
mush[(Mushroom Observations)] --> prep
prep[Prepare Data] --> train
train[Train AI Model] --> calc
train --> model@{ shape: lean-l , label: AI Model}
calc[Calculate Model Predictions] --> present
present[Present Results]
Metrics and time series visulizations are currently hosted as a Streamlit dashboard, so check it out.
For a more detailed overview, see either Issues or the detailed Project Planner, where the project is managed.