[Calc] Uber greener than TfL?
mdjmay4 opened this issue · comments
The Uber / TfL balance seems off in my history - not a dev though so struggle to work out exactly why. E.g. One trip I did last month was TfL one way (on the tube, 37min, 1.7kg), and Uber back (25min, 10.12km, 1.2kg). This seems wrong to me, without being able to see the calcs in app I can't verify though.
Hi @mdjmay4. Thank you so much for taking the time to give us feedback, it's extremely valuable.
When looking at an activity in the app there's a section with links to the references behind the calculations. Is this what you're looking for?
Hi @sorensvejstrup !
Sort of - but as I understand those are just the references for how it might be calculated (which is important), not how it actually is being calculated by the app (e.g. 35 mins x 20 km/h x 0.1 kg/km = 1.17kg, or whatever it is).
My question is more about the veracity of the end result, as I'd be surprised if Uber was more efficient - but without seeing the calculations it's hard to understand what's going on.
Hey, the calculations for different transportation modes are available in co2eq/transportation.js - but I agree that it does sound suspicious, so if you can send me a screenshot of the debugging information for each activity I'd like to look into it :)
How to get debugging information
- go to profile --> settings (gear cog in upper right corner)
- under "Advanced Settings", enable "Debug mode"
- open the activity and scroll to the bottom
Great - attached below!
Thanks for sharing the screenshots. We looked into it and our current estimation of average car emissions is too low, so we're updating the numbers now. Thanks for pointing it out! 💪
Great - no worries! TfL themselves report much lower numbers for public transport as well (my journey would be ~0.5kg according to this but I understand methodologies will differ.
Edit: I think it's because the durationToDistance assumes a much healthier average speed than what we get here in London... 🙂
Haha yeah, that could very well be it. And we also can't make it specific to the type of transport used (e.g. bus, train, tube), so we're going with an average estimation for public transport based on the this source. Hopefully we can get more precise along the way :)
Closing this for now - model should now be slightly improved. Thanks @mdjmay4 for raising this!