Example of SMHI direct use

Direct usage of SMHI.

Basic use

The SMHI client is in an experimental state and currently only deals with Metobs data. It is intended as a parent class for easy traversing of SMHI observational data from GPS or city names directly.

The most basic usage of the client is to simply ask for data for a given station and a given parameter:

from smhi.smhi import SMHI

client = SMHI()

client.parameters.data #List all available parameters
#Parameter 1 is hourly air temperature

stations = client.get_stations(1)
stations.data #List all available stations for parameter 1

data = client.get_data(1, 72630) #Get data from specific station
#Station 72630 is Gothenburg

The data variable holds four dataframes when a single station is used to fetch data as above. They are called station, parameter, period and df, where the latter holds the actual observational data. See further example of Metobs use.

Interpolation

When there are several stations close to a measurement location and historical data records are incomplete, it is convenient to use nearby stations to fill in the missing data: This naive interpolation feature is included in the package by the parameter radius. Thus, if we want to fetch historical data on the snow coverage in Kiruna, the call is:

from smhi.smhi import SMHI

client = SMHI()

#Parameter 8 is depth of snow coverage
data = client.get_data(8, 180960, 40) #Get data from specific station
#Station 180960 is Kiruna. 40 indicates we will use stations within
#a 40km radius to complement any data losses.

data2 = client.get_data(8, 180960) #Get comparison data without the
#interpolation.

Visualise the data:

Scatter plot code

import plotly.graph_objects as go

d1 = data.df
d2 = data2.df

index = d1.index.intersection(d2.index)
d2_dropped = d2.drop(index, axis=0)

fig = go.Figure()
fig.add_trace(
    go.Scattergl(
        x=d1.index,
        y=d1["Snödjup"],
        mode="markers",
        name="Kiruna station"
    )
)
fig.add_trace(
    go.Scattergl(
        x=d2_dropped.index,
        y=d2_dropped["Snödjup"],
        mode="markers",
        name="Interpolerat, radie 40 km"
    )
)
fig.update_layout(
    title='Historiskt snödjup i Kiruna',
    xaxis_title="År",
    yaxis_title="Snödjup [m]",
    legend={"orientation": "h"},
    margin={"l": 0, "r": 0, "b": 80, "t": 100}
)

fig.show()

Finding data from a city

There are a lot of stations available: Frequently, we are simply interested in meterological observations from a specific city. Using the library geopy, this is included in the package too:

from smhi.smhi import SMHI

client = SMHI()

#Parameter 8 is depth of snow coverage
data = client.get_data_by_city(8, "Kiruna", 40) #Get data from a station
#close to the city centre of Kiruna, and fill any empty holes with data
#from stations within a 40km perimeter