Python DS Workshop - Pandas and Pyplot

import numpy as np import pandas as pd

athletes = pd.read_csv("Athletes.csv")

# Pandas DataFrame athletes

# Pandas Series, you can treat this like an array athletes['Discipline']

athletes[['Discipline']]

athletes[['Discipline', 'Name']]

athletes['NOC']

athletes['NOC'] == 'Egypt'

athletes[athletes['NOC'] == 'Egypt']

len(athletes[athletes['NOC'] == 'Egypt'])

athletes.loc[10:21,['Name', 'Discipline']]

athletes.iloc[10:21,:1]

import matplotlib.pyplot as plt

medals = pd.read_csv('Medals.csv') medals

medals['Not Gold'] = medals['Silver'] + medals['Bronze']

plt.scatter(medals['Gold'], medals['Not Gold']) plt.xlim(0, 40) plt.ylim(0, 80) plt.xlabel("Gold Medals") plt.ylabel("Other Medals Medals") plt.title("Bronze + Silver vs. Gold Medals per country")