In February 2020 I made a summary of our household energy use, going back to 2013, with all of the usage normalized into kWh. I wanted to repeat the process for 2020, but unfortunately I hadn't taken any notes on how I had gathered the data the first time. So I ended up reverse engineering my methodology from the data.


Our electricity is billed through PSE. From their website, under "my usage", there is an option to download a CSV with historical usage data. You can choose to export all billing data, which will give a CSV with monthly aggregated kWh and costs, or you can provide a range of dates and get daily aggregated energy usage without costs. My existing data appears to have come from the first option (and I'd like to have the costs information), so I'm using that again. I had to strip out the dollar sign from the cost column (using the Text to Column feature in LibreOffice, with $ as the deliminator), but otherwise the new rows just dropped into my existing spreadsheet.

Total electric usage in 2020 was 5 MWh, down 20% year-over-year. The pricing is tiered based on usage, plus taxes, less credits. But all in all, our average cost per kWh remained at $0.09, although it had dropped to $0.08 between June 2019 and March 2020.

Natural Gas

Natural gas usage was downloaded together with the electricity usage from PSE. Gas is metered in CCF, but billed in therms using a BTU factor that changes throughout the year depending on the fuel mix. 1 therm is 100k BTU or about 29.3 kWh, which is the conversion factor I use to normalize gas usage in kWh.

Our gas usage was 12 MWh, which was a 15% reduction year-over-year. The average cost per kWh remained $0.03, which has been consistent for back 2013 (except a brief increase in 2015 Q2).


Gasoline usage isn't nearly as simple, particularly since I stopped tracking gallons, mileage, and price per gallon years ago. I do, however, track what we spend on fuel for each vehicle. What I had done last year was estimate the gallons of fuel based on fuel expenses and the average price per gallon in Washington. Gas prices in our area runs a bit higher than the average for the state, so I increased the average price per gallon by 6%. This gave me an average of 13.5 gallons per fill up in the Jeep and 10 gallons for the Matrix, which seems about right.

The problem was... I couldn't remember where I got the average fuel price data. GasBuddy has a wonderful chart, but no APIs or data tables. I stuck the column name of my price per gallon table ("Weekly Washington Regular All Formulations Retail Gasoline Prices (Dollars per Gallon)") into DuckDuckGo and it looks like I got the data from the US Energy Information Administration.

I suppose that verbose column name was worth something after all. I confirmed that several rows in my existing data matched the EIA website, then copied the new rows in. "Text to Column" was useful again to help Calc interpret the date format.

The fuel expenses themselves were exported from GnuCash where they were already separated by vehicle. I exported the transactions as a CSV and then ran the output through grep and awk to make it match format in my spreadsheet.

grep Expenses: fuel-expenses.csv | awk -F, '{ OFS=","; print "Gasoline", $11, substr($1, 7, 4), $1, $4, $13}' | grep 2020

To convert gallons of gasoline to kWh, I'm using a constant of 33.41 kWh/gal. This conversion factor came from the Gasoline gallon equivalent. This puts us at 659 gallons or 22 MWh of energy use for 2020. This is a 40% reduction year-over-year, which is not surprising since I stopped commuting in March.


Like gasoline, I am tracking propane usage based our transaction history in GnuCash. Propane is only used in the BBQ and we only swapped out the propane tank once in 2020. To normalize to kWh, I assume 18 lbs of propane per tank and a conversion factor of 22000*0.2931/1000 = 6.4482 kWh/lb. I'll have to reverse engineer how I arrived at this conversion later... Propane use was down 50% year-over-year, but we don't go through much propane even in a normal year.


Overall, our total energy usage was down 33% year-over-year. The drop in gasoline was expected, since I wasn't commuting, but the reduction in both power and natural gas was quite unexpected. Perhaps it helped that I was home to turn off the lights when April left during the day. Or maybe it had something to do with reduced diaper laundry after Anders potty trained.

I'd like to do the same thing with water usage, which might help confirm/disprove theories about why power/gas usage dropped. I'd also like to move the data from a spreadsheet to a database and build some charts of the monthly data to make it easier to see trends.