from forecast import Forecast
weather = Forecast()
# weather.clean_tmy('../data/golden.csv')
weather.load_tmy('../data/golden_clean.csv')
weather.add_predictor('Dry-bulb (C)').add_predictor('Dew-point (C)')
predictions = weather.auto_regressive(12, 2)
print(predictions)
weather.persist()
n10 = n10DB[hour]
n10.rename('n10', inplace=True)
n50 = n50DB[hour]
n50.rename('n50', inplace=True)
n100 = n100DB[hour]
n100.rename('n100', inplace=True)
data = pandas.concat([n10, n50, n100], axis=1)
p = seaborn.boxplot(data=data)
p.set(title='Dry bulb temperature ranges at 15:00 on July 1st', xlabel='Number of scenarios generated', ylabel='Dry Bulb (C)')
seaborn.plt.show()
"""
predictions, test = atlanta.auto_regressive(48, 50, trim_data=False)
db = predictions[atlanta.predictor_variables[0]]
testAtl = test[atlanta.predictor_variables[0]]
predictions, test = golden.auto_regressive(48, 50, trim_data=False)
goldenDB = predictions[golden.predictor_variables[0]]
testGold = test[golden.predictor_variables[0]]
print('...predictions succeeded, now plotting...')
f, (ax, ax2) = plot.subplots(2,1)
p = seaborn.boxplot(data=goldenDB, ax=ax)
p2 = seaborn.boxplot(data=db, ax=ax2)
seaborn.tsplot(testGold, ax=ax)
seaborn.tsplot(testAtl, ax=ax2)
#ax.set_title('GLobal Horizontal Irradiation (W/m^2)')
ax.set_title('Dry-Bulb (C) for Golden, CO')
# Create the forecast object (default horizon = 24 hours):
weather = Forecast() # set horizon with Forecast(horizon=48)
# load weather data:
#weather.clean_tmy('../data/houston.csv') # this saves a 'cleaned' tmy file
weather.load_tmy('../data/atlanta_clean.csv')
# set the simulation time if necessary:
weather.simulation_time = '07/01/1900 01:00'
# Set up the VAR model by adding variables:
weather.add_predictor('Dry-bulb (C)').add_predictor('RHum (%)')
# Make a number of predictions and capture the predictions and the test data:
predictions, test = weather.auto_regressive(72, 50, trim_data=False)
# Extract a variable of interest (index values correspond to the order the variables were added):
db = predictions[weather.predictor_variables[0]]
rh = predictions[weather.predictor_variables[1]]
# ...and the test data:
dbtest = test[weather.predictor_variables[0]]
rhtest = test[weather.predictor_variables[1]]
# use seaborn to make a pretty plot:
f, (ax, ax2) = plot.subplots(2, 1)
pretty = seaborn.boxplot(data=db, ax=ax)
seaborn.tsplot(dbtest, ax=ax)
pretty2 = seaborn.boxplot(data=rh, ax=ax2)
