pythonpandasnumpybisect

I want to change all of the number in my array into ABCD scoring system using python


I want to change all the number in my array into ABCD scoring system. I have 100 X 1 array of scores, my array is put under total_score. I tried using

def grade(score):
    if 91 <= score <= 100:
        return 'A'
    if 81 <= score <= 90.99:
        return 'B'
    if 71 <= score <= 80.99:
        return'C'
    if 61 <= score <= 70.99:
        return'D'
    else:
        return'E'
grade = float(total_score)

I also tried using


def determine_grade(scores, breakpoints=[50, 60, 70, 80, 90], grades='FEDCBA'):
    i = bisect.bisect(breakpoints, scores)
    return grades[i]
    [grade(score) for score in [total_score]]

both doesnt work. There are also few suggestion, but it seems none of the suggestion work with pandas array


Solution

  • Let's say your total scores array looks like this:

    total_scores = ['99', '100', '52', '69', '33', '77']
    

    Your grading function:

    def grade(score):
        if 91 <= score <= 100:
            return 'A'
        if 81 <= score <= 90.99:
            return 'B'
        if 71 <= score <= 80.99:
            return'C'
        if 61 <= score <= 70.99:
            return'D'
        else:
            return'E'
    

    Then all you have to do is:

    grading_score = [grade(float(score)) for score in total_scores]  # returns ['A', 'A', 'E', 'D', 'E', 'C'] 
    

    Note that I'm converting the strings in my array into floats before calling the 'grade' function.

    Here is the official documentation for using list comprehensions