take a dump of Github JSON issues data on https://github.com/webcompat/web-bugs/ See issue https://github.com/webcompat/webcompat.com/issues/1233

issues_stats.py

import json

RESULTS = '''
new format: {new}
old format: {old} # can't conclude anything
---------------------
INVALID: On {invalid} > empty: {empty_invalid}
VALID:   On {valid} > empty: {empty_valid}
'''


def body_trim(body):
    '''Extract only the string starting with Expected Behavior.'''
    search = body.partition('**Steps to Reproduce**')[2]
    search = search.replace('\n', '').replace('\r', '')
    return search


def is_empty(reduced_body):
    '''Check if the issue has an empty description.'''
    new_format = True
    emptiness = False
    if 'Expected Behavior' in reduced_body:
        # More chances to be without description
        if 'Expected Behavior:Actual Behavior:' in reduced_body:
            desc = reduced_body.partition('Expected Behavior')
            # but the description can be after actual behavior
            if desc[2] == ':Actual Behavior:':
                emptiness = True
            # or the description can be in the steps to reproduce
            if desc[0].endswith('2) \xe2\x80\xa6'):
                emptiness = True
        else:
            emptiness = False
    else:
        new_format = False
        emptiness = False
    return emptiness, new_format


with open('issues.json') as issuesdb:
    issues = json.load(issuesdb)

old_counter, new_counter = 0, 0
invalid, valid = 0, 0
empty_invalid, empty_valid = 0, 0

for issue in issues:
    status_list = [label['name'] for label in issue['labels']
                   if label['name'].startswith('status-')]
    # We just need the part of the body which starts after Steps to reproduce
    body = issue['body'].encode('utf-8')
    reduced_body = body_trim(body)
    # Checking for invalid issues only
    if 'status-invalid' in status_list:
        invalid += 1
        invalid_empty, new_format = is_empty(reduced_body)
        if invalid_empty:
            empty_invalid += 1
    # bug is considered valid
    else:
        valid += 1
        valid_empty, new_format = is_empty(reduced_body)
        if valid_empty:
            empty_valid += 1
    if new_format:
        new_counter += 1
    else:
        old_counter += 1

print(RESULTS.format(
    new=new_counter, old=old_counter,
    invalid=invalid, valid=valid,
    empty_invalid=empty_invalid, empty_valid=empty_valid))