mlproject script

mlproject.py

import pandas as pd
import numpy as np
import pandas as pd
import io
import requests

ign_url='https://gist.githubusercontent.com/zhonglism/09e35db741cd0ad2ac5162bb02d246ae/raw/e9f45068bfb80d2be0264045d652110a79ffe092/ign.csv'
vgsale_url="https://gist.githubusercontent.com/zhonglism/f146a9423e2c975de8d03c26451f841e/raw/f79e190df4225caed58bf360d8e20a9fa872b4ac/vgsales.csv"

ign_con=requests.get(ign_url).content
vgsale_con=requests.get(vgsale_url).content

ign=pd.read_csv(io.StringIO(ign_con.decode('utf-8')))
vgsales=pd.read_csv(io.StringIO(vgsale_con.decode('utf-8')))


# ign=pd.read_csv('ign.csv')
# vgsales=pd.read_csv('vgsales.csv')

ign=ign.drop_duplicates()
vgsales=vgsales.drop_duplicates()
ign=ign.drop('Unnamed: 0',axis=1) 

#remove platforms which has less than 10 titles
ign_platform=ign.groupby(['platform'])['title'].count()
ign_platform=pd.DataFrame(ign_platform[ign_platform<10])
ign_platform=ign_platform.index.tolist()

for platform in ign_platform:
    ign=ign[ign['platform']!=platform]

vgsales_platform=vgsales.groupby(['Platform'])['Name'].count()
vgsales_platform=pd.DataFrame(vgsales_platform[vgsales_platform<10])
vgsales_platform=vgsales_platform.index.tolist()

for pf in vgsales_platform:
    vgsales=vgsales[vgsales['Platform']!=pf]


find_replace={'Dreamcast':'DC','Game Boy':'GB','Game Boy Advance':'GBA','GameCube':'GC',
 'Genesis':'GEN','NES':'NES','Nintendo 3DS':'3DS','Nintendo 64':'N64','Nintendo DS':'DS',
 'Nintendo DSi':'DS','Playstation':'PS','PlayStation 2':'PS2','PlayStation 3':'PS3',
 'PlayStation 4':'PS4','PlayStation Portable':'PSP','PlayStation Vita':'PSV','Super NES':'SNES',
 'Wii U':'WiiU','Xbox':'XB','Xbox 360':'X360','Xbox One':'Xone'}

for key,value in find_replace.items():
    ign=ign.replace({key:value})


result=ign.merge(vgsales,left_on=['title','platform'], right_on=['Name','Platform'], how='inner')
result[['title','platform','Name','Platform']].tail(5)

Here is comment