Project idea: Analysis of weather data¶

The topic of this project revolves around the analysis of weather data given by an open-source provider Bright Sky.

How to proceed?¶

1. Use the provided function get_weather_data to receive a Pandas DataFrame.
2. Plot relevant measures and inspect the results.
3. Try using a clustering algorithm to distinguish factors. For instance, are there distinct months where sunshine is low and high?
4. Try different dates and assess how

In [ ]:

# Add you packages here
%pip install pandas

# Add you packages here %pip install pandas

In [ ]:

import requests
import pandas as pd

from typing import Optional
from datetime import date, datetime

import requests import pandas as pd from typing import Optional from datetime import date, datetime

In [6]:

def get_weather_data(
    since_date: str,
    end_date: Optional[str] = None,
    dwd_station_id: str = "01766",
) -> pd.DataFrame:
    """Fetched weather data from the Bright Sky API.

    Example:
        >>> df = get_weather_data(since_date="2017-01-01")
        >>> df.head()

    Args:
      since_date (str):     The date to start from. Format is YYYY-MM-DD.
      end_date (str):       The date to end at. Format is YYYY-MM-DD.
                            If not specificed, will default to today.
      dwd_station_id (str): The Deutscher Wetterdienst ID of
                            the station of measurement.

    Returns:
      pd.DataFrame: A Pandas DataFrame containing the weather data.
    """

    validate_date(since_date, "since_datw")
    url = "https://api.brightsky.dev/weather"

    if not end_date:
        # If no end date is provided, use today's
        end_date = date.today()
    else:
        # If provided, check if it correct
        validate_date(end_date, "end_date")

    params = {
        "last_date": date.today(),
        "date": date.fromisoformat(since_date),
        "dwd_station_id": "01766",
    }

    headers = {"Accept": "application/json"}

    response = requests.get(url, headers=headers, params=params)
    response.raise_for_status()

    df = pd.DataFrame(response.json()["weather"])

    # Parse the timestamps
    df["timestamp"] = pd.to_datetime(df["timestamp"])

    return df


def validate_date(date: str, date_type: str):
    """Validates the format of a date to comply with YYYY-MM-DD"""
    try:
        # We need to make sure that the correct date format is used
        datetime.strptime(date, "%Y-%m-%d")
    except ValueError:
        raise ValueError(
            f"The date format supplied for `{date_type}` is of the wrong format.",
            "Please use YYYY-MM-DD to use this function",
        )

def get_weather_data( since_date: str, end_date: Optional[str] = None, dwd_station_id: str = "01766", ) -> pd.DataFrame: """Fetched weather data from the Bright Sky API. Example: >>> df = get_weather_data(since_date="2017-01-01") >>> df.head() Args: since_date (str): The date to start from. Format is YYYY-MM-DD. end_date (str): The date to end at. Format is YYYY-MM-DD. If not specificed, will default to today. dwd_station_id (str): The Deutscher Wetterdienst ID of the station of measurement. Returns: pd.DataFrame: A Pandas DataFrame containing the weather data. """ validate_date(since_date, "since_datw") url = "https://api.brightsky.dev/weather" if not end_date: # If no end date is provided, use today's end_date = date.today() else: # If provided, check if it correct validate_date(end_date, "end_date") params = { "last_date": date.today(), "date": date.fromisoformat(since_date), "dwd_station_id": "01766", } headers = {"Accept": "application/json"} response = requests.get(url, headers=headers, params=params) response.raise_for_status() df = pd.DataFrame(response.json()["weather"]) # Parse the timestamps df["timestamp"] = pd.to_datetime(df["timestamp"]) return df def validate_date(date: str, date_type: str): """Validates the format of a date to comply with YYYY-MM-DD""" try: # We need to make sure that the correct date format is used datetime.strptime(date, "%Y-%m-%d") except ValueError: raise ValueError( f"The date format supplied for `{date_type}` is of the wrong format.", "Please use YYYY-MM-DD to use this function", )

In [5]:

# Example usage
df = get_weather_data("2017-01-01")
df.head()

# Example usage df = get_weather_data("2017-01-01") df.head()

Out[5]:

	timestamp	source_id	precipitation	pressure_msl	sunshine	temperature	wind_direction	wind_speed	cloud_cover	dew_point	relative_humidity	visibility	wind_gust_direction	wind_gust_speed	condition	precipitation_probability	precipitation_probability_6h	solar	fallback_source_ids	icon
0	2017-01-01 00:00:00+00:00	7003	0.0	1024.59	NaN	2.2	230.0	11.5	87.0	-2.4	72.0	15000.0	220.0	18.4	dry	None	None	0.0	{}	cloudy
1	2017-01-01 01:00:00+00:00	7003	0.0	1023.70	NaN	1.7	210.0	9.4	62.0	-2.4	74.0	15000.0	200.0	17.3	dry	None	None	0.0	{}	partly-cloudy-night
2	2017-01-01 02:00:00+00:00	7003	0.0	1023.30	NaN	0.8	230.0	9.0	50.0	-2.4	79.0	15000.0	210.0	21.6	dry	None	None	0.0	{}	partly-cloudy-night
3	2017-01-01 03:00:00+00:00	7003	0.0	1022.30	0.0	1.0	200.0	13.0	87.0	-1.6	83.0	12000.0	180.0	22.3	dry	None	None	0.0	NaN	cloudy
4	2017-01-01 04:00:00+00:00	7003	0.0	1021.80	0.0	0.2	190.0	11.9	87.0	-1.7	87.0	8000.0	160.0	20.5	rain	None	None	0.0	NaN	cloudy