Fetch historical weather data with Python using various packages

historical_weather.py

# Fetch historical weather data with Python using various packages

# use plotnine for graphing
from plotnine import *
import pandas as pd
import pathlib


# pynasapower
# Fetches NASA POWER data

from pynasapower.get_data import query_power
from pynasapower.geometry import point, bbox
import datetime

# Get weather data for the IRRI Zeigler Experiment Station
gpoint = point(121.255669, 14.16742, "EPSG:4326")
start = datetime.date(2000, 1, 1)
end = datetime.date(2001, 12, 31)
irri_data = query_power(geometry = gpoint,
                    start = start,
                    end = end,
                    to_file = False,
                    community = "ag",
                    parameters = ['T2M', 'RH2M', 'PRECTOTCORR'],
                    temporal_api = "daily",
                    spatial_api = "point",
                    format = "csv")

irri_data

# melt the data so we can visualise all three parameters

irri_data = pd.melt(irri_data, id_vars=['YEAR', 'DOY'],
                    value_vars=['T2M','RH2M','PRECTOTCORR'])

# use plotnine to visualise the data
(
    ggplot(irri_data, aes("DOY", "value"))
    + geom_line(aes(colour = "factor(YEAR)"))
    + facet_wrap("variable")
)

# Fetch weather station data

# meteostat
from datetime import datetime
from meteostat import Point, Daily
from meteostat import Stations

stations = Stations()
stations = stations.nearby(lat=14.16742, lon=121.255669)
station = stations.fetch(10)

station

# Ninoy Aquino Airport, WMO 98429, appears to be the best choice for
#  long-term weather data and is hourly

# Create point for IRRI (note that these are reversed from above)

# Set time period
start = datetime(2000, 1, 1)
end = datetime(2001, 12, 31)

# Get daily data for 2000 and 2001
naia_data = Daily('98429', start, end)
naia_data = naia_data.fetch()

# The naia_data are incomplete, missing many daily values highlighting
# one of the issues about good weather data from weather stations in
# many important locations
naia_data

# melt the naia_data to visualise it, note that there is no RH in this set
naia_data.index.name = 'date'
naia_data.reset_index(inplace=True)
naia_data['date'] = pd.to_datetime(naia_data['date'])
naia_data['doy'] = naia_data['date'].dt.dayofyear
naia_data['year'] = naia_data['date'].dt.year


naia_data = pd.melt(naia_data,
                    id_vars=['year','doy'],
                    value_vars=['tavg','prcp'])

# Use plotnine again to plot line chart including average, minimum and maximum temperature

(
    ggplot(naia_data, aes("doy", "value"))
    + geom_line(aes(colour = "factor(year)"))
    + facet_wrap("variable")
)


# tsgettoolbox fetches many weather data sets including GHCN and GSOD
# I ended up using GSODR to identify stations closest to IRRI to specify
# them for download here as there was no clear way to do that with this
# module.

# This module doesn't appear to work with the modified example from the
# README

from tsgettoolbox import tsgettoolbox
df = tsgettoolbox.ncei_gsod(sites="02329500", startDT="2000-01-01")

# gsodpy
# This module doesn't appear to work with the modified example from the
# README

import gsodpy

gdown.get_data(directory="noaa_gsod")

print(getmembers(gsodpy), isfunction)

# open-meteo
# I couldn't get any of the clients to work with this API properly to
# fetch weather data.