Quick analysis of EMF 2014 talks

Results

Total Entries =  184
Errors        =   20
Used Entries  =  164
Min           =   15 min
Max           =  180 min
Average       =   52 min
----------------------------------------
Duration (min) | Count
            30 |  48
            60 |  38
            45 |  34
            90 |  19
            15 |  14
           120 |  10
           180 |   1

talk_analysis.py

import json
from datetime import datetime
from collections import OrderedDict

DATETIME_FORMAT = '%Y-%m-%dT%H:%M:%SZ'

def get_events(file_name):
    f = open(file_name, 'r')
    return json.load(f)['conference_events']['events']


def get_length_in_minutes(event):
    end = event.get('end_time', None)
    start = event.get('start_time', None)
    if end is None or start is None:
        return None
    end = datetime.strptime(end, DATETIME_FORMAT)
    start = datetime.strptime(start, DATETIME_FORMAT)
    duration = end - start
    return duration.seconds / 60


def get_hist(entries):
    hist = {}
    for e in entries:
        hist[e] = 1 if e not in hist else hist[e] + 1

    return OrderedDict(sorted(hist.items(), key=lambda x: x[1], reverse=True))


def main(file_name):
    events = get_events(file_name)

    all_durations = [get_length_in_minutes(e) for e in events]

    errors = len([d for d in all_durations if d is None])
    durations = [d for d in all_durations if d is not None]

    print('Total Entries = {0: >4d}'.format(len(all_durations)))
    print('Errors        = {0: >4d}'.format(errors))
    print('Used Entries  = {0: >4d}'.format(len(durations)))
    print('Min           = {0: >4d} min'.format(int(min(durations))))
    print('Max           = {0: >4d} min'.format(int(max(durations))))
    print('Average       = {0: >4d} min'.format(int((sum(durations) / len(durations)))))
    print('-' * 40)
    hist = get_hist(durations)

    print('Duration (min) | Count')
    for time, count in hist.items():
        print('{0: >14d} | {1: >3d}'.format(int(time), count))


if __name__ == '__main__':
    talks_2014 = '2014/events.json'
    main(talks_2014)