import pandas as pd
import numpy as np
from energaze.models import PostRegression
# Load your data
df = pd.read_csv('energy_data.csv')
# Initialize model with data
model = PostRegression(
data=df,
consumption_var='consumption',
treatment_var='treatment',
time_var='date',
site_var='site_id',
treatment_date='2023-01-01'
)
# Fit the model
model.fit()
# Get treatment effects
results = model.get_treatment_effect()
print(f"Treatment Effect: {results['effect']:.2f}")
print(f"P-value: {results['p_value']:.4f}")
# Get detailed summary
summary = model.summary()
print(summary)
# State-level grouping
state_model = PostRegression(
data=df,
consumption_var='consumption',
treatment_var='treatment',
time_var='date',
site_var='site_id',
state_var='state', # Group by state
treatment_date='2023-01-01'
)
state_model.fit()
state_results = state_model.get_treatment_effect()
# Results for each state
for state, result in state_results.items():
print(f"{state}: Effect = {result['effect']:.2f}")
# Combined State × Wave grouping
combined_model = PostRegression(
data=df,
consumption_var='consumption',
treatment_var='treatment',
time_var='date',
site_var='site_id',
state_var='state',
wave_var='wave', # Both state and wave
treatment_date='2023-01-01'
)
from energaze.models import TWFE, TimePeriodSavingsRegression
# Compare different models
models = {
'PostRegression': PostRegression(
data=df, consumption_var='consumption',
treatment_var='treatment', time_var='date',
site_var='site_id', treatment_date='2023-01-01'
),
'TWFE': TWFE(
data=df, consumption_var='consumption',
treatment_var='treatment', time_var='date',
site_var='site_id', treatment_date='2023-01-01'
),
'TimePeriodSavings': TimePeriodSavingsRegression(
data=df, consumption_var='consumption',
treatment_var='treatment', time_var='date',
site_var='site_id', treatment_date='2023-01-01'
)
}
# Fit all models and compare
results = {}
for name, model in models.items():
model.fit()
results[name] = model.get_treatment_effect()
print(f"{name}: Effect = {results[name]['effect']:.2f}")
