Meters data features

There is a data set for each kind of meter: electricity, hotwater, chilledwater, steam, water, irrigation, solar and gas. Each data set contains the timestamp in the format YYYY-MM-DD hh:mm:ss (local timezone) and one column per building with the meter reading in KWh. There is a reading per hour. In meters folder there are three subfolders:

• Raw
• Screening
• Cleaned

Raw

As the name says, these are the raw data sets, the original data for each meter.

Screening

Anomalies

This folder contains a file for each meter, with atypical values marked. Outliers in the raw meters data set were detected using the Seasonal Hybrid ESD (S-H-ESD) developed by Twitter. This part was implemented in R language, the process can be found here.

Breakout

This folder contains a file for each meter, with marked values where a breakout is detected. For breakouts detection in the cleaned meters data set was used the Breakout Detection package developed by Twitter, choosing 168 points (a week) as minimum to define a gap. A brief introduction about this package can be found here. Cleaned meter data (outliers and 24hs zero readings removed) are used for this analysis. This was implemented in R language, the process can be found here.

Cleaned

This folder contains all cleaned data sets: outliers and whole-day zero readings removed. In this notebook can be found the process to create them.