A way to find mode values within range

Question

I have an numpy array of dimension values in this structure:

arr = array([[3067,   78, 3172,  134],
             [3237,   89, 3394,  128],
             [3475,   87, 3743,  141],
             [3763,   86, 3922,  131],
             [3238,  147, 3259,  154]])

which basically stores the location of data located in the screen, where the values represented as: [x_left, y_top, x_right, y_bottom]]

I only need to work on the x_left values, as I am trying to find where on the page I am most likely able to find these objects.

I am aware of scipy.mode, which returns mode values. Is there a way to return multiple modes, say the top 10 mode values in a given numpy column? Better yet, is there a way to use mode so that a mode is within a given range? For example, the lines above have x_left values of 3237 and 3238, which are fairly closely aligned. Is there a way to calculate these two as a single mode value?

fmarm · Accepted Answer · 2019-10-31 00:31:31Z

1

You can convert your numpy array column to a pandas series and use .value_counts()

import pandas as pd
x_left = pd.Series(arr[:,0])
x_left.value_counts()
#3475    1
#3237    1
#3067    1
#3763    1
#3238    1
#dtype: int64

You could also round the values to, for example the nearest 10 integer to group values between ranges

def customRound(x, base=10):
    return base * round(x/base)
x_left_round = x_left.apply(customRound)
x_left_round.value_counts()
#3240    2
#3760    1
#3070    1
#3480    1
#dtype: int64

You can then see that you have two x_left values close to 3240

answered Oct 31, 2019 at 0:31

fmarm

4,2741 gold badge20 silver badges30 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

A way to find mode values within range

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related