With the addition of including the context of the weather for historical consumption and generation data, having historical weather is also important. One low cost way to get this weather data is OpenWeatherMap which offers a 5-year JSON file of timestamped weather conditions at hourly resolution for many cities in the U.S. for $10 currently.

https://openweathermap.org/history-bulk

This capability of being able to purchase a bulk weather set and be able to use that weather data to analyse the solar generation or power consumption time series in the context of that weather is useful. SolarQuant is one engine that could use this weather upload for testing and verification of its data-fitting algorithms.