proc_w3gadswpy.py

#!pip install html5lib #install html5lib, only needs to be run once
#You might need to restart kernel after running with the menu Kernel>Restart
import pandas as pd
import numpy as np
import urllib

#description and prices of the Xeon Gold processors
df_xeon_golds=pd.read_html('https://proxy.mentoracademy.org/getContentFromUrl/?userid=brooks&url=https://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors', header=0)[78]
#statistics about the performance of a range of Intel processors from cpu-monkey
df_stats=pd.read_csv('https://proxy.mentoracademy.org/getContentFromUrl/?userid=brooks&url=https://gist.github.com/cab938/6499da85d31cfccc9cc5b13621963312/raw/34db3b55bd14f39fc59e6b5128b667a9061f77d7/cpu_performance.csv')

#clean up the price column in df_xeon_golds
def clean_price(price):
    try:
        if str(price).startswith('$'):
            return int( str(price)[1:])
        else:
            return None
    except:
        return None

df_xeon_golds["price"]=df_xeon_golds["Release price (USD)"].apply(clean_price)

joined_df=pd.merge(df_xeon_golds, df_stats, left_on=["Model number"], right_on=["processor"], how="left")

def calculate_stats(row):
    if not np.isnan(row['price']) and not np.isnan(row['performance']):
        #higher is better value for the money
        row['price_performance']=row['performance']/row['price']
    return row

joined_df=joined_df.apply(calculate_stats, axis=1) #apply the function above to create a new column with price per performance
joined_df=joined_df[joined_df['price_performance'].notnull()] #just consider those processors which have stats
joined_df['price_performance_rank']=joined_df['price_performance'].rank(ascending=False) #determine the ranks for each processor in a new column
joined_df.sort_values(by='price_performance_rank').head(5) #print out the top five items