findperc.py

def find_perc(data, bin_name):
    # find percentatge for each type of effect
    sum_wt = data.s006.sum()
    # tax liability decrease of > $500
    large_cut = tc.zsum(data.s006[data['change'] < -500]) / float(sum_wt)
    # tax liability decrease of $100-500
    small_cut = tc.zsum(data.s006[(data['change'] >= -500) & (data['change'] <= -100)]) / float(sum_wt)
    # tax liability change of less than $100
    no_change = tc.zsum(data.s006[(data['change'] > -100) & (data['change'] < 100)]) / float(sum_wt)
    # tax liability increase of $100-500
    small_increase = tc.zsum(data.s006[(data['change'] >= 100) & (data['change'] <= 500)]) / float(sum_wt)
    # tax liability increase of > $500
    large_increase = tc.zsum(data.s006[(data['change'] > 500)]) / float(sum_wt)
    
    # find edges for the bar chart
    bottom = 0.0
    total_perc = bottom
    edge_name_list = [bottom, large_cut, small_cut, no_change, small_increase, large_increase]
    edges_list = []
    for item in edge_name_list:
        total_perc += item
        edges_list.append(total_perc)
    plot_data = pd.DataFrame({'bottom': [edges_list[0]],
                              'large_cut': [edges_list[1]],
                              'small_cut': [edges_list[2]],
                              'no_change': [edges_list[3]],
                              'small_inc': [edges_list[4]],
                              'large_inc': [edges_list[5]],
                              'name': [bin_name]})
    return ColumnDataSource(plot_data)