Useful link: Top 10 weather APIs

Data sources

cities_table = {"berlin": 1, "hamburg": 2, "muenchen": 3, "koeln": 4, "frankfurt": 5, "stuttgart": 6, "bremen" :7, "leipzig": 8, "hannover": 9, "nuernberg": 10, "dortmund": 11, "dresden": 12, "kassel": 13, "kiel": 14, "bielefeld": 15, "saarbruecken": 16, "rostock": 17, "freiburg": 18, "magdeburg": 19, "erfurt": 20}

provider_list = { 'timeanddate_com': 0, 'wetter_com': 1, 'owm': 2, 'wetter_de': 3, 'accuweather': 4, 'wunderground': 5, 'historical': 6}

Potential weather web sites (and their data source if known)

Which of the following are now working on the server?

openweathermap.org + API - [Claus]
wetter.com (ecmwf) - [Jan]
timeanddate.com/weather (customweather) - [Alicja]
wetter.de (meteogroup) - [Greg]
wunderground + API - [Ines]
accuweather [Denis]
(weather.com [Claudia] - abandoned)
(yahoo weather + API - abandoned)
(wetter24.de (meteogroup) - abandoned)
(wetter.info (meteogroup + DWD) - abandoned)

FYI: Weather.com and Wunderground need to be accessed via their APIs. We did not succeed in scraping them. They do however share the same parent company and now use a single API scheme for data access (despite having different data).

variables of interest

air temperature
precipitation
humidity
air pressure (?)
wind speed
wind direction
hours of sunshine
cloud cover

Data representation

Structure of the dictionary:
{'site' : (String) e.g. 'Wetter.de'
'city' : (String)
'date' : int
'prediction_time': int

'hourly': {'00' (String) : { *see 'hourly' below }
'01' (String) : { ... }
...
'24' (String) : { ... }

'daily': {'0' (String) : { *see 'daily' below }
: {'1' (String): {...}
...
: {'10' (): {...}
}

for each hour:
'hourly' {'00' : {
'temp' : Celsius
'humidity' : %
'pressure' : hPa
'wind_speed': m/s
'rain_amt': mm = liter/m²
'rain_chance': %
'cloud_cover': %
}

for each day:
'daily': {'0': {
'high': Celsius
'low': Celsius
'rain_chance': %
'rain_amt': mm = liter/m²
'pressure': hPa (ave)
'cloud_cover': % (ave)
}

Datatypes: