1 Millions of movies

import pandas as pd import seaborn as sns import matplotlib.pyplot as plt import numpy as np

sample_data = False size = 1000

titles_df = pd.read_csv('/work/data/ImdbTitleBasics.csv', low_memory = True) #names_df = pd.read_csv('ImdbTitleCrew.csv') #names_df = pd.read_csv('ImdbTitleRatings.csv')

names_df = pd.read_csv('/work/data/ImdbName.csv') principals_df = pd.read_csv('/work/data/ImdbTitlePrincipals.csv')

titles_df.head(5)

names_df.head(5)

principals_df.head(5)

df = principals_df.merge(titles_df, left_on='tconst', right_on='tconst').merge(names_df, left_on= 'nconst', right_on='nconst')

if sample_data: df = df.sample(n=size)

df.head()

df.describe(include='all')

df = df.drop(['characters','primaryTitle'], axis=1)

df['endYear'].describe()

df['endYear'].isnull().sum()

df = df.replace("\\N", np.nan)

df['endYear'].isnull().sum()

df.isnull().sum()

df['job'].head(20)

df = df.drop(['job'], axis=1)

df = df.drop(['primaryProfession'], axis=1)

df[df['endYear'].notnull()]

df = df[df['runtimeMinutes'].notnull()] df = df[df['genres'].notnull()]

df.isnull().sum()

df = df.convert_dtypes()

df.dtypes

df.head()

df.nunique()

df_genres = df['genres'].str.split(",", expand=True)

df_genres[0].unique()

df = df.drop(['genres'], axis=1) df = df.drop(['knownForTitles'], axis=1)

df_clean = pd.concat([df, df_genres], axis=1)

df_clean[['category', 'titleType','isAdult']] = df_clean[['category', 'titleType','isAdult']].astype('category')

df_clean['runtimeMinutes'] = df_clean['runtimeMinutes'].astype('int')

df_clean[['runtimeMinutes','startYear','endYear','birthYear','deathYear']] = df_clean[['runtimeMinutes','startYear','endYear','birthYear','deathYear']].astype('int', errors='ignore')

df_clean.head()

df_clean.value_counts(['category']).plot(kind='barh')

df_clean.value_counts(['titleType']).plot(kind='barh')

df_clean.value_counts(['primaryName'])[:10].plot(kind='barh')

df_actors = df_clean[df_clean.category.isin(['actor', 'actress'])]

df_actors['startYear'].astype('int', errors='ignore')

fig, ax = plt.subplots(figsize=(40,10)) b = sns.pointplot(x=0, y="startYear", data=df_clean, ax=ax) b.axes.set_title("Category per Year",fontsize=50) b.set_xlabel("Categories",fontsize=50) b.set_ylabel("Year",fontsize=50) b.tick_params(labelsize=20)