Stacked bar chart for two data sets, normalized by the number of points within each bin, with numpy and matplotlib

normalized_stacked_bars.ipynb

stacked_bars.py

def stacked(adata, bdata, x_min, x_max, bin_width, title, xlabel):
    # by Cassie Lumbrazo, 7/29/2022
    
    np.seterr(invalid='ignore') #ignore the runtime errors for 0/0
    
    # define datasets 
    a = adata.copy(deep=True)
    b = bdata.copy(deep=True)

    # note that these give you bin edges, not bin centers
    a_count, a_bin_edges = np.histogram(a, bins=range(int(x_min), int(x_max), int(bin_width)))
    b_count, b_bin_edges = np.histogram(b, bins=range(int(x_min), int(x_max), int(bin_width)))
    
    # to get bin centers, add 1/2 of the bin width (in this case 1/2 of 1 is 0.5) to the bin edges 
    # (but drop the last bin edge value)
    half_range = bin_width / 2
    a_bin_centers = a_bin_edges[:-1] + half_range # SINCE SW IN 100 w/m2 BINS
    b_bin_centers = b_bin_edges[:-1] + half_range

    # calculate percent based on total (a and b) in each bin
    a_percent = a_count / (a_count + b_count)
    b_percent = b_count / (a_count + b_count)

    # remove nans where there was missing a or b data
    a_percent[np.isnan(a_percent)] = 0
    b_percent[np.isnan(b_percent)] = 0

    # make the stacked bar plot
    fig, ax = plt.subplots()
    ax.bar(a_bin_edges[1:], a_percent, width=bin_width, label=namesnow, color=colorsnow, alpha=.7)
    ax.bar(a_bin_edges[1:], b_percent, width=bin_width, bottom=a_percent, label=namesnowunload, color=colorsnowunload, alpha=.7)
    plt.xlim(x_min, x_max)
    plt.title(title)
    plt.xlabel(xlabel)
    plt.grid()
    plt.legend()