The following code gives me a general centroid of all of the readings of my df.
pos = df4[['x', 'y']].to_numpy() # gives me all the x-,y-coordinates in df4
def centroid(arr):
length = arr.shape[0]
sum_x = np.sum(arr[:, 0])
sum_y = np.sum(arr[:, 1])
return sum_x/length, sum_y/length
coll_cps = np.array(centroid(pos)) # Create centroids between ids
How can I create a new column with temporary centroids of each person ID for, lets say, every 10th reading?
My df looks like this:
x y id time
0 162 282 2700 0
1 162 282 2819 0
2 162 282 2820 0
3 449 235 2700 1
4 449 235 2820 1
5 449 235 2819 1
6 457 293 2819 2
7 457 293 2820 2
8 457 293 2700 2
9 164 283 2700 3
10 164 283 2819 3
11 164 283 2820 3
12 457 293 2700 4
13 457 293 2820 4
14 457 293 2819 4
15 450 235 2700 5
16 450 235 2820 5
17 450 235 2819 5
18 449 234 2700 6
19 449 234 2819 6
20 449 234 2820 6
21 456 293 2820 7
22 456 293 2819 7
23 456 293 2700 7
24 167 277 2820 8
25 167 277 2700 8
26 167 277 2819 8
27 167 277 2820 9
28 167 277 2700 9
29 167 277 2819 9
... ... ... ...
The output should be a new column with the temporary centroids between the id's within x rows, 10 for instance. So the average centroid per 10 readings at a time.
So, for 10 rows at a time, append the average centroid for each id.
Pop in a helper column to identify group and then use the groupby and apply pattern as such:
import pandas as pd
# some data
x_vals = [1, 2, 3, 10, 11, 20]
y_vals = [2, 4, 6, 0, 10, 0]
data = {'x': x_vals, 'y': y_vals}
df = pd.DataFrame(data)
group_size = 3
# make "helper row" with group number
df['group'] = df.index//group_size
def centroid(row):
return (row.x.mean(), row.y.mean())
df_centroids = df.groupby('group').apply(centroid)
print(df)
print()
print(df_centroids)
Yields:
x y group
0 1 2 0
1 2 4 0
2 3 6 0
3 10 0 1
4 11 10 1
5 20 0 1
group
0 (2.0, 4.0)
1 (13.666666666666666, 3.3333333333333335)