西电数据挖掘大作业之电影评级数据分析

# -*- coding: utf-8 -*- import pandas as pd import numpy as np mnames = ['movieid', 'title', 'genres'] movie = pd.read_table('movies.dat', sep='::', header=None, names=mnames, engine='python') print('----------------------------------------------------------------------------') print("examples of movies:\n") print(movie.head(5)) print('\nthe length of movies data:%d' % len(movie)) print('----------------------------------------------------------------------------\n') unames = ['userid', 'gender', 'age', 'occupation', 'zip'] user = pd.read_table('users.dat', sep='::', header=None, names=unames, engine='python') print('----------------------------------------------------------------------------') print("examples of users:\n") print(user.head(5)) print('\nthe length of users data:%d ' % len(user)) print('----------------------------------------------------------------------------\n') rnames = ['userid', 'movieid', 'ratings', 'timestamp'] rating = pd.read_table('ratings.dat', sep='::', header=None, names=rnames, engine='python') print('----------------------------------------------------------------------------') print("examples of ratings:\n") print(rating.head(5)) print('\nthe length of ratings data:%d ' % len(rating)) print('----------------------------------------------------------------------------\n') # 把三张表的数据合并到一起 data = rating.merge(user).merge(movie) print('----------------------------------------------------------------------------') print("examples of datas:\n") print(data.head(5)) print('----------------------------------------------------------------------------\n') # 按性别计算每部电影的平均得分 meanratings = pd.pivot_table(data, index=['title'], values='ratings', columns=[ 'gender'], aggfunc=np.mean, fill_value=0) print('----------------------------------------------------------------------------') print('The average score of each film (by Gender):') print(meanratings[:10]) print('----------------------------------------------------------------------------\n') # 过滤掉评分数据不到100条的电影,再按年龄统计电影的平均得分 movie_rating = data.groupby(['title']).size() # print(movie_rating[:10]) movie_ix = movie_rating.index[movie_rating > 100] mrating = meanratings.loc[movie_ix] # print(mrating.head()) # 女性最喜欢的电影 print('----------------------------------------------------------------------------') print('Women favorite movies TOP10: ') print(mrating.sort_values(by=['F'], ascending=False).head(10)) print('----------------------------------------------------------------------------\n') print('----------------------------------------------------------------------------') print('Men favorite movies TOP10: ') print(mrating.sort_values(by=['M'], ascending=False).head(10)) print('----------------------------------------------------------------------------\n') # 找出男性和女性评分分歧最大的电影 # 按男女评分差值的绝对值来统计 mrating['diff'] = np.abs(mrating['F'] - mrating['M']) #print(mrating.sort_values(by=['diff'], ascending=False).head(5)) # 不考虑性别因素，分歧最大的电影（用方差来统计） rating_std = data.groupby(by=['title'])['ratings'].std() rating_std_n = rating_std.loc[movie_ix] # 按照前面超过100条评论的电影进行过滤 # print(rating_std_n.head()) rating_std_n.fillna(0).sort_values(ascending=False)[:10] # 查看评分最多的电影,最热门电影 ratings_by_title = data.groupby('title').size() print('----------------------------------------------------------------------------') print('The hottest movie (the most rated movie TOP10):') print(ratings_by_title.sort_values(ascending=False).head(10)) print('----------------------------------------------------------------------------\n') # 查看最高分电影 # 先算出每部电影平均得分 mean_ratings = data.pivot_table( values='ratings', index='title', aggfunc='mean') # 对电影平均得分排序 print('---------------------------------------------------------------------------------') print('High score movie TOP20') print(mean_ratings.sort_values(ascending=False).head(20)) print('----------------------------------------------------------------------------------\n') # 查看评分最多且评分最高电影 # 选出评分次数大于1000的电影 hot_movies = ratings_by_title[ratings_by_title > 1000] # print(len(hot_movies)) # print(hot_movies.head(10)) # 将这些电影的评分进行排序 hot_movies_rating = mean_ratings[hot_movies.index] top_10_good_movies = hot_movies_rating.sort_values(ascending=False) print('--------------------------------------------------------------------------------------------') print('Good movie TOP10(high score and large score):') print(top_10_good_movies.head(10)) print('--------------------------------------------------------------------------------------------\n')