2. Scrape and Plot 2015 Baseball Standings (matplotlib)

Begin by importing our modules



In [1]:

    
import pandas as pd
import matplotlib.pyplot as plt
import ipywidgets
from IPython.display import display
from collections import OrderedDict

%matplotlib inline

pd.options.mode.chained_assignment = None  # default='warn'

Now load the master data we'll use



In [2]:

    
dfMaster = pd.read_csv('../data/dayByDayStandings2015.csv')
dfMaster.fillna(0, inplace=True)
dfMaster['date'] = pd.to_datetime(dfMaster['date'])
dfMaster['date'] = dfMaster['date'].apply(pd.datetools.normalize_date)
dfMaster.set_index(['division', 'date'], inplace=True)
dfMaster = dfMaster[['Tm', 'W', 'L', 'GB', 'RS', 'RA', 'W-L%', 'pythW-L%']]

Review the data

First look at the metadata



In [3]:

    
dfMaster.info()









    



<class 'pandas.core.frame.DataFrame'>
MultiIndex: 5460 entries, (AL East, 2015-04-05 00:00:00) to (NL West, 2015-10-03 00:00:00)
Data columns (total 8 columns):
Tm          5460 non-null object
W           5460 non-null int64
L           5460 non-null int64
GB          5460 non-null object
RS          5460 non-null int64
RA          5460 non-null int64
W-L%        5460 non-null float64
pythW-L%    5460 non-null float64
dtypes: float64(2), int64(4), object(2)
memory usage: 358.8+ KB

Now examine the data itself



In [4]:

    
dfMaster.tail()

Make a plot - the hard way

Also the intuititive way



In [5]:

    
dfALCentral = dfMaster.ix['AL Central', ['Tm', 'GB']]



In [6]:

    
dfALCentral.replace({'--': 0}, inplace=True)
dfALCentral['GB'] = dfALCentral['GB'].astype(float)



In [7]:

    
cle = dfALCentral[dfALCentral.Tm == 'CLE']
kcr = dfALCentral[dfALCentral.Tm == 'KCR']
minn = dfALCentral[dfALCentral.Tm == 'MIN']
chw = dfALCentral[dfALCentral.Tm == 'CHW']
det = dfALCentral[dfALCentral.Tm == 'DET']
plt.figure()
plt.plot(cle.index.values, cle['GB'], label='CLE')
plt.plot(kcr.index.values, kcr['GB'], label='KCR')
plt.plot(minn.index.values, minn['GB'], label='MIN')
plt.plot(chw.index.values, chw['GB'], label='CHW')
plt.plot(det.index.values, det['GB'], label='DET')
plt.legend(loc='best')
plt.show()

What's wrong here?

Y-scale is inverted.
Too small and crowded.
Colors are wrong.
Only covers one division - there are five others.

How to fix?

Invert the Y values prior to plotting.
Set the figure size manually - don't rely on defaults.
Make the plot dynamic so you can switch divisions on the fly.



In [8]:

    
dropList = OrderedDict(zip(dfMaster.index.levels[0].values, dfMaster.index.levels[0].values))

mlb_colors = {'HOU': '#072854', 'CIN': '#C6011F', 'NYM': '#FB4F14', 'PHI': '#BA0C2F',
              'LAD': '#083C6B', 'LAA': '#B71234', 'COL': '#333366', 'TOR': '#003DA5', 
              'WSN': '#BA122B', 'BAL': '#ED4C09', 'STL': '#C41E3A', 'SDP': '#002147', 
              'ARI': '#A71930', 'MIL': '#92754C', 'MIN': '#C6011F', 'MIA': '#F9423A', 
              'BOS': '#C60C30', 'OAK': '#003831', 'PIT': '#FDB829', 'CHC': '#003279', 
              'CHW': '#000000', 'SFG': '#F2552C', 'DET': '#001742', 'TBR': '#00285D', 
              'KCR': '#15317E', 'ATL': '#B71234', 'TEX': '#BD1021', 'SEA': '#005C5C', 
              'CLE': '#003366', 'NYY': '#1C2841'}

Pass the division into a plotting function.
Plot each series with a for loop.
Since it's a timeline, extend the X-axis - make it a rectangle.
Remove axes



In [11]:

    
def plot_division(division):
    tmpMaster = dfMaster.ix[division, ['Tm', 'GB']]

    plt.figure(figsize=(15,6))
    ax = plt.subplot(111)
    ax.spines["top"].set_visible(False)  
    ax.spines["bottom"].set_visible(False)  
    ax.spines["right"].set_visible(False)  
    ax.spines["left"].set_visible(False)
    for team in enumerate(tmpMaster.Tm.value_counts().index.values):
        tmp = tmpMaster[tmpMaster.Tm == team[1]]
        tmp.replace({'--': 0}, inplace=True)
        tmp['GB'] = tmp['GB'].astype(float)
        tmp.loc[:,'GB'] = tmp.loc[:,'GB'] * -1
        plt.plot(tmp.index.values, tmp['GB'], label=team[1], color=mlb_colors[team[1]])

    plt.legend(loc='best')
    plt.show()


divW = ipywidgets.Dropdown(options=dropList)
init = divW.value
j = ipywidgets.interactive(plot_division, division=divW)
display(j)



In [ ]:

		Tm	W	L	GB	RS	RA	W-L%	pythW-L%
division	date
NL West	2015-10-03	LAD	91	70	--	661	592	0.565	0.550
	2015-10-03	SFG	84	77	7.0	693	620	0.522	0.551
	2015-10-03	ARI	78	83	13.0	715	710	0.484	0.503
	2015-10-03	SDP	74	87	17.0	647	725	0.460	0.448
	2015-10-03	COL	67	94	24.0	730	841	0.416	0.436

2. Scrape and Plot 2015 Baseball Standings (matplotlib)

Begin by importing our modules

Now load the master data we'll use

Review the data

First look at the metadata

Now examine the data itself

Make a plot - the hard way

What's wrong here?

How to fix?

1. Create some metadata for a dropdown list and colormap.

2. Embed a dropdown in the notebook that handles the plotting for us.