Download only the period you need
PG&E provides an example of customer tools offering 15-minute electricity data and downloadable usage records. Availability and format differ by utility and account. Use your provider’s own export instructions and keep the original file unchanged as a reference.
Choose a period tied to a question, such as a week with an unusual overnight pattern. You do not need to share the file publicly to perform a basic local comparison.
Distinguish energy from average power
If an interval contains 0.25 kWh over 15 minutes, average power during that interval is 0.25 ÷ 0.25 = 1 kW. It is not 0.25 kW. Conversely, a column already labeled average kW should not be added as though it contained kWh.
For a hypothetical day containing 96 complete quarter-hour rows totaling 8 kWh, daily energy is 8 kWh. Check timestamp conventions and daylight-saving transitions before assuming every calendar day must contain exactly 96 rows.
Mark patterns before naming appliances
Note repeated overnight consumption, cooking-time peaks and differences between occupied and away days. Compare those times with your own activity notes. Several appliances can run in the same interval, and a short high-power event may be averaged with quieter minutes.
A utility chart does not by itself prove a faulty refrigerator or unauthorized electricity use. Treat unexplained patterns as questions for further observation, not diagnoses.
Reconcile before calculating cost
Check that interval totals reasonably match the billed period, allowing for stated corrections, missing data and boundary times. Identify whether the file describes imports, exports or another quantity if solar or storage is present. For a time-of-use estimate, apply the correct rate to each applicable period; one average rate loses that detail. Save your assumptions beside the analysis so another month can be compared consistently.
Choose your next step
Continue with the question that matches your next decision:
Sources & further reading
Numerical examples are illustrative unless explicitly described as measured. This guide does not claim hands-on product testing. Read the methodology and editorial disclosure.