Loop through a list of car numbers to calculate mean of steering angle for each car

I am working on correcting the steering angle. Steps: 1) Filter the data based on backstretchend and backstretchstart of a race track 2) Calculate the mean of existing values by per car 3) Calculate the difference between the mean calculated per car and the actual values of steering in collected data 4) Calculate the mean of this difference

I have filtered and have calculated mean on overall column

Sample data:

car lap steering
00   20  10
00   21 -12
00   22  8
00   21  9
1    20  2
1    21  3
1    21 -9
1    22  8

Code:

backstretchstart = 0.5
backstretchend = 0.59
#Step 1: Filter the data based on backstretchend and backstretchstart of a race track 
df = df.query('corrected_lap__frac > 0.50 & corrected_lap__frac < 0.59')
avgsteerback = df["steering"].mean()
df["steering_correct"] = df["steering"] - avgsteerback
avgsteer = df["steering_correct"].mean()
if avgsteer <= 0 and avgsteer != (np.isnan(avgsteer)):
    df["steering_correct"] = df["steering_correct"] * - 1