Rating distribution

%matplotlib inline
import pandas as pd
import matplotlib.pyplot as plt

import sys
from pathlib import Path
root = Path(".").absolute().parent.parent
sys.path.append(str(root))

import caption_contest_data as ccd

def get_dist_plot(contest, alg="KLUCB"):
    df = ccd.summary(contest)

    ax = df.score.hist(bins=20, grid=False)
    ax.set_xlim(1, 3)
    ax.set_xticks([1, 2, 3])
    ax.set_xticklabels(["Unfunny", "Somewhat funny", "Funny"])
    ax.set_title(f"Ratings for contest={contest}")
    ax.set_yscale("log")
    return df, ax

df, ax = get_dist_plot(538)

df, ax = get_dist_plot(640)

df, ax = get_dist_plot(586)

df.head()

	rank	funny	somewhat_funny	unfunny	count	score	precision	contest	caption
0	1	2575	2327	1783	6685	2.118474	0.009769	586	Of course it looks alien to you. You've never ...
1	2	2164	2448	1500	6112	2.108639	0.009806	586	I have a feeling they are going to treat us li...
2	3	1771	1543	1474	4788	2.062030	0.011865	586	Left on Pennsylvania Avenue, big white house o...
3	4	1901	2088	1595	5584	2.054799	0.010564	586	And they say life can't exist in a vacuum.
4	5	1114	1221	957	3292	2.047691	0.013801	586	Wow. Aliens really do perform the jobs most Am...

df = ccd.summary(580)

idx = df.funny == 0
idx.sum() / len(idx)

0.4646220719632939

idx = df.somewhat_funny == 0
idx &= df.funny == 0
idx.sum() / len(idx)

0.1357160106254528