Replies: 2 comments
-
Hi, I'm not sure how to respond here without writing half a book. In general, if I knew the answers to most of your questions, they'd be documented somewhere. In fact, no one here really knows the details about what you are asking. I'll pick out a few specifics that I do want to try and answer.
In general, you should always remember that was not originally intended for technical analysis. It was (is) intended for live visualization where delays of a few 100ms do not matter as they are barely noticeable by eye. The raw session time is something like packet timestamps with an unknown delay that varies for each packet. These can be corrected partially (in the case of timing data). I can't really comment on your idea for merging and 'EstDistance'. Given the low sample rate of the data and the existing inaccuracies, things that should work in theory often do not work well in practice. Therefore, trial and error is often the only way forward. |
Beta Was this translation helpful? Give feedback.
-
Yeah, it was a bit of an essay! Thanks for the response! As you say, the intent of the data is not for detailed analyses, so I probably shouldn't be surprised there are so many unknowns. Given how widespread the use of the data and your tools are, it's a shame there's not more official/inside knowledge around. I would guess that media and the teams do get more explanations (and potentially higher accuracy data). I would also assume that F1 are aware of a lot of the X, YouTube and Medium analyses using this data but it's not in their interests to support further. Picking up on your bullets:
I'll post again with how I get on. There's a big risk that I make everything look good with my various "corrections" but actually degrade the accuracy. Fingers crossed! |
Beta Was this translation helpful? Give feedback.
-
Hi all, I'm newish to F1 (DTS generation) and very new to FastF1 and GitHub (first post), so please bear with me and beware my assumptions! I've been playing around with FastF1 for the last few weeks and have some thoughts/questions.
First off, thank you, @theOehrly!
The data F1 provides is great but there are also clearly gaps and inaccuracies. I would love feedback on my assumptions regarding Car and Position Data.
Car Data (get_car_data)
RPM
Speed
nGear
Throttle, Brake and DRS
SessionTime
Position Data (get_pos_data)
X, Y and Z
SessionTime
Approach for merging car and position data
Thanks and congrats if you read all/much of that! Let me know what you’ve found, your thoughts on my assumptions and any ideas on approaches.
Beta Was this translation helpful? Give feedback.
All reactions