plots.py 文件源码

def plot_cluster_hist(X_featurized, labels, num_clusters):
    '''
    Plot histograms of users and corresponding clusters.

    Parameters
    ----------

    X_featurized : array-like
        Featurized Data

    labels: array-like
        Predicted cluster to data.

    num_clusters: int
        Number of clusters.

    Returns
    -------

    Plot : matplotlib.lines.Line2D
        Figure.

    '''

    fig = plt.figure()
    ax_ = fig.add_subplot(1,1,1)

    # Set colors.
    # Create DataFrame with features and labels.
    # Note sklearn cluster naming starts at zero, so adding 1 is convenient.

    X_featurized['label'] = labels + 1
    # Parameters for plotting.
    params_ = {'ax': ax_ , 'bins': np.arange(num_clusters +2) - 0.5}
    # Plot cluster and corresponding color.
    X_featurized.label.plot(kind = 'hist', **params_)

    # Format figure.

    ax_.set_title("Number of users in each cluster.", fontsize =14, fontweight='bold')
    ax_.set_xticks(range(1, num_clusters +1))
    ax_.set_xlim([0, num_clusters + 1])
    ax_.set_ylim([0,1200])
    ax_.set_xlabel('Cluster')
    ax_.set_ylabel("Number of users")

    # plt.savefig('cluster_hist')
    plt.show()