nws_demo.sagews

Using CoCalc to Look at National Weather Service Hourly Forecast

A short demo written as the first blizzard of the season arrived in Chicago.

Here's a picture from the NW suburbs the next morning.

References

• NWS

  • hourly weather graphs

  • REST interface

• Python package noaa-sdk

  • noaa-sdk on pypi

  • noaa-sdk on github

• Chicago area blizzard forecast in the news, evening of 2018-11-25

  • Chicago Sun-Timeshttps://chicago.suntimes.com/news/nws-issues-winter-storm-watch-for-sunday-more-than-6-inches-snow-possible/)

  • Daily Herald

Package note

The following step is no longer needed. As of 2018-11-28, noaa-sdk is globally installed on CoCalc.

If noaa_sdk isn't available globally in CoCalc yet, open a .term and do the following. Then restart your worksheet.

pip3 install --user noaa_sdk

%auto
%default_mode python3

my_zip_code = '60004'

from noaa_sdk import noaa
n = noaa.NOAA()

# get hourly forecast for next 48 hours
res = n.get_forecasts(postal_code=my_zip_code, country='US', hourly=True)[:48]

# display first three rows of data from NWS
print(res[:3])

[{'number': 1, 'name': '', 'startTime': '2018-11-26T15:00:00-06:00', 'endTime': '2018-11-26T16:00:00-06:00', 'isDaytime': True, 'temperature': 28, 'temperatureUnit': 'F', 'temperatureTrend': None, 'windSpeed': '15 mph', 'windDirection': 'NW', 'icon': 'https://api.weather.gov/icons/land/day/bkn?size=small', 'shortForecast': 'Mostly Cloudy', 'detailedForecast': ''}, {'number': 2, 'name': '', 'startTime': '2018-11-26T16:00:00-06:00', 'endTime': '2018-11-26T17:00:00-06:00', 'isDaytime': True, 'temperature': 26, 'temperatureUnit': 'F', 'temperatureTrend': None, 'windSpeed': '15 mph', 'windDirection': 'WNW', 'icon': 'https://api.weather.gov/icons/land/day/bkn?size=small', 'shortForecast': 'Mostly Cloudy', 'detailedForecast': ''}, {'number': 3, 'name': '', 'startTime': '2018-11-26T17:00:00-06:00', 'endTime': '2018-11-26T18:00:00-06:00', 'isDaytime': True, 'temperature': 24, 'temperatureUnit': 'F', 'temperatureTrend': None, 'windSpeed': '15 mph', 'windDirection': 'WNW', 'icon': 'https://api.weather.gov/icons/land/day/bkn?size=small', 'shortForecast': 'Mostly Cloudy', 'detailedForecast': ''}]

# get ready to use dataframes
import pandas as pd

# convert short forecast wording to numeric level
# only order matters: 0, 1, 2
def snow_number(s):
    if 'Heavy Snow' in s:
        ret = 2
    elif 'Snow' in s:
        ret = 1
    else:
        ret = 0
    return ret

# put NWS data into a pandas dataframe
df = pd.DataFrame(res)[['startTime','temperature','shortForecast']]
# fix types
df['Time'] = pd.to_datetime(df['startTime'])
df['Snowfall'] = df['shortForecast'].apply(snow_number)
df = df.set_index('Time')
df.head()

	startTime	temperature	shortForecast	Snowfall
Time
2018-11-26 21:00:00	2018-11-26T15:00:00-06:00	28	Mostly Cloudy	0
2018-11-26 22:00:00	2018-11-26T16:00:00-06:00	26	Mostly Cloudy	0
2018-11-26 23:00:00	2018-11-26T17:00:00-06:00	24	Mostly Cloudy	0
2018-11-27 00:00:00	2018-11-26T18:00:00-06:00	23	Mostly Cloudy	0
2018-11-27 01:00:00	2018-11-26T19:00:00-06:00	23	Mostly Cloudy	0

# plot forecast for temperature and snowfall
# key to snowfall:
#   2 = heavy
#   1 = chance of snow
#   0 = no snow

df[['temperature','Snowfall']].plot(subplots=True, sharex=True, figsize=(6,2.5), fontsize=8);