sciencelee

sciencelee

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

df = pd.read_csv("vgsales.csv")

df = df.loc[df['Year'] < 2008]
df.head(3)
pd.set_option('display.max_columns', None)

sciencelee

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

df = pd.read_csv("vgsales.csv")

sciencelee

#  If this was SQL, it would look something like this:
'''
SELECT Year, SUM Global_Sales
FROM df
GROUP BY Year
'''

# In Python...
# Grouping it by year produces a DataFrameGroupBy object
df.groupby(['Year'])

sciencelee

'''
SELECT SUM Global_Sales
FROM df
GROUP BY Publisher
'''

# Grouped by Publisher and selected the sum of Global_Sales (a Series object)
df.groupby(['Publisher'])['Global_Sales'].sum() 

# Now we should be able to plot it??

sciencelee

# We can make this into a horizontal bar graph to make this a little easier to read.
df.groupby(['Publisher'])['Global_Sales'].sum().plot(kind='barh')

# That last line produces a messy bar graph with hundreds of publishers.  
# We can use iloc or head mehthods to limit the number shown
df.groupby(['Publisher'])['Global_Sales'].sum().head(10).plot(kind='barh') 

sciencelee

# Flatten the grouped object as seen above
df_flat = df.groupby(['Publisher'])['Global_Sales'].sum().reset_index()

# sort it by Global_Sales, include only the first 10, and make a horizontal bar plot.  Yay!
df_flat.sort_values(['Global_Sales'], ascending=False).head(10).plot(kind='barh', y='Global_Sales', x='Publisher')

sciencelee

# SQL Query might look like...
'''
SELECT Year, SUM Global_Sales
FROM df_me
GROUP BY Year, Platform
'''

# plot global sales by platform by year.  Sounds tricky, but we can handle it.

platforms = ['NES', 'PS', '2600', 'PS2', 'GBA']

sciencelee

# This line of code does it for us
df_me.groupby(['Year', 'Platform'])['Global_Sales'].sum().unstack().plot()  

sciencelee

import pandas as pd
df = pd.read_csv('kc_house_data.csv')