ggplot

import pandas as pd
import random
import matplotlib as pl
import pylab
from pandas import DataFrame
from ggplot import *


def flipTrials(minExp, maxExp, numTrials):
    """Assumes minExp and maxExp positive integers; minExp < maxExp
    Plots results of 2**minExp to 2**maxExp coin flips"""

    cols = ['x', 'ratio_means', 'ratio_stdvs', 'ratio_coeffvar',
            'diffs_means', 'diffs_stdvs', 'diffs_coeffvar']
    df = DataFrame(index=range(minExp, maxExp + 1), columns=cols)
    for i in df.index:
        numFlips = 2**i
        numHeads = np.random.randint(0, 2, [numTrials, numFlips]).sum(1)
        numTails = np.ones(numTrials)*numFlips - numHeads
        ratio = numHeads / numTails
        diffs = abs(numHeads - numTails)
        df.x[i] = numFlips
        df.ratio_means[i] = ratio.mean()
        df.ratio_stdvs[i] = ratio.std()
        if df.ratio_means[i] > 0:
            df.ratio_coeffvar[i] = df.ratio_stdvs[i] / df.ratio_means[i]
        df.diffs_means[i] = diffs.mean()
        df.diffs_stdvs[i] = diffs.std()
        if df.diffs_means[i] > 0:
            df.diffs_coeffvar[i] = df.diffs_stdvs[i] / df.diffs_means[i]

    # for faceting create a dataframe with all columns stacked and indexed by colname
    dfm = pd.melt(df[['x', 'ratio_means', 'ratio_stdvs', 'diffs_means', 'diffs_stdvs',
                     'ratio_coeffvar', 'diffs_coeffvar']], id_vars='x')
    print ( ggplot(dfm, aes(x='x', y='value', colour='variable'))
            +geom_line() + scale_x_log() + facet_wrap('variable') )
    return df, dfm

random.seed(0)
df, dfm = flipTrials(2, 10, 2)
print 'Test:\n ', df[-5:]