#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Thu Mar 15 13:11:42 2018

@author: teacher
"""

import pandas as pd
import matplotlib.pyplot as plt

# Read csv file in as a dataframe
movies = pd.read_csv("imdb_1000.csv")

# Plot histogram of duration of all movies
movies['duration'].plot.hist()

action_cond = movies['genre'] == 'Action'
action_movies = movies[action_cond]

crime_movies = movies[movies['genre'] == 'Crime']

action_mean = action_movies['duration'].mean()
crime_mean = crime_movies['duration'].mean()
print("Mean of duration of action movies",action_mean)
print("Mean of duration of crime movies",crime_mean)

plt.show()
action_movies['duration'].plot.hist(alpha = 0.5, bins=20)
crime_movies['duration'].plot.hist(alpha = 0.5, bins = 20)

