Make CO2 calculations based on liters about sustainable-mobility-api

[MarkusShepherd]

That sounds very interesting and useful to me. I see two ways to calculate the CO2 emission by litre of fuel:

1. Obtaining the numbers of emission by volume of fuel. In many respects that should be easy and fairly accurate as that's the very source of CO2 being emitted.
2. If we use average km/l or miles/gallons numbers for cars, we can convert the fuel volume into distance travelled, and plug this into our model as-is.

[brylie]

@ErikM1995, thanks for the suggestion. I am taking a more in-depth look into this idea now. I will post my findings as comments here so that we can discuss things like the data sources/model and API.

[MarkusShepherd]

Yeah, more or less. I'd think analogous to Mode we could create a class Fuel, an Enum with different instances like PETROL, DIESEL, LPG (?), and field avg_co2_per_liter.

As for the method API, I don't like these overly long names (just couldn't be bothered to read all that stuff), so would rather have something like estimator.from_distance() and estimator.from_fuel() or similar. Not a hill I'm willing to die on though... 😉

[brylie]

Also for starters, I was planning to use these metrics:

https://www.nrcan.gc.ca/sites/www.nrcan.gc.ca/files/oee/pdf/transportation/fuel-efficient-technologies/autosmart_factsheet_6_e.pdf

[brylie]

@MarkusShepherd in terms of Python architecture, would we perhaps create a new class like CO2_Estimator that handles mode- and fuel-based estimations?

[brylie]

Relatedly, Transport and Environment published an interactive article today called "How clean are electric cars?".

To answer this question we have developed a tool (see below) that compiles all the most up-to-date data on CO2 emissions linked to the use of an electric, diesel or petrol car.

We might be able to dig into their source data for our purposes.

[MarkusShepherd]

@MarkusShepherd in terms of Python architecture, would we perhaps create a new class like CO2_Estimator that handles mode- and fuel-based estimations?

What would that class look like? I'm not sure I see a benefit in a class that just throws different purposes together.

If we want to have one unified way to calculate emissions (and we'd first need to convince ourselves that's something useful), the best way I see is to change the estimator.estimate_co2() signature to

def estimate_co2(**kwargs) -> float:
    # do something

and then select the lower level functionality depending on what arguments were passed. But right now I'd say it'll only obscure the API, so if distance and fuel based estimates are the only two use cases, why not just have two different methods or classes to call?

[brylie]

So for now, just defining a new function like the following?

def estimate_co2_from_fuel_consumed(fuel_type, volume, units):
    # howdy

[brylie]

I fully agree with everything you say above, which is why I am asking you about the API design ;-)

[MarkusShepherd]

Hmmm... I've stumbled across the question of units for fuel calculations. Some are measured by volume (litres, gallons, cubic metres, ...), while others are measured by mass (kg, tonne, whatever Americans use). Then there's electricity of course that's measure directly in energy... You noted the unit in some example above, but at that point I only thought of litres vs gallons, not volume vs mass.

So we would need co2_per_unit and unit for each fuel type, plus potentially some conversions between units. Also, it would be great to link Mode and Fuel via litres/km (or miles/gallon for the metrically challenged 😜).

Sorry, turned out a bit of rambling, seems like there's quite a number of use cases that need to be considered...

[brylie]

OK. Do you know of a good Python library we can lean on for unit conversion?

[brylie]

@ErikM1995 we have a pull request open (#43) now that will hopefully fulfil your feature request! :-)

[ErikM1995]

Wow, nice work guys! Great to see that you really took the time to think about the possibilities of this idea. About the units of measurement, I always use https://www.co2emissiefactoren.nl/lijst-emissiefactoren/ . This is a pretty good databank formed by the Dutch ministry of economics and environment and some energy companies, but unfortunately that's in Dutch. Many of the fuel names are in English though.

They also open up the debate whether you'd want to incorporate the well to wheel, well to tank or tank to wheel CO2 ratio, interesting stuff!