Python seaborn 模块，stripplot() 实例源码

def plot_similardishes(idx,xlim):
    match = yum_ingr2.iloc[yum_cos[idx].argsort()[-21:-1]][::-1]
    newidx = match.index.get_values()
    match['cosine'] = yum_cos[idx][newidx]
    match['rank'] = range(1,1+len(newidx))

    label1, label2 =[],[]
    for i in match.index:
        label1.append(match.ix[i,'cuisine'])
        label2.append(match.ix[i,'recipeName'])

    fig = plt.figure(figsize=(10,10))
    ax = sns.stripplot(y='rank', x='cosine', data=match, jitter=0.05,
                       hue='cuisine',size=15,orient="h")
    ax.set_title(yum_ingr2.ix[idx,'recipeName']+'('+yum_ingr2.ix[idx,'cuisine']+')',fontsize=18)
    ax.set_xlabel('Flavor cosine similarity',fontsize=18)
    ax.set_ylabel('Rank',fontsize=18)
    ax.yaxis.grid(color='white')
    ax.xaxis.grid(color='white')

    for label, y,x, in zip(label2, match['rank'],match['cosine']):
         ax.text(x+0.001,y-1,label, ha = 'left')
    ax.legend(loc = 'lower right',prop={'size':14})
    ax.set_ylim([20,-1])
    ax.set_xlim(xlim)

def plot_null(fnull, fcore, ax=None):
    """
    Plot the null shared response overlayed with the actual number of
    shared responders.
    """
    null = pd.read_csv(fnull, sep='\t')
    core = pd.read_csv(fcore, sep='\t', index_col=0)
    ncore = core.groupby('overall').size()

    for i in [-1, 0, 1]:
        if i not in ncore:
            ncore.loc[i] = 0

    order = ['health', 'mixed', 'disease']

    if ax is None:
        fig, ax = plt.subplots()

    sns.stripplot(data=null, x='type', y='n', order=order, jitter=True, ax=ax, alpha=0.2)

    ax.scatter([0, 1, 2], [ncore.loc[-1], ncore.loc[0], ncore.loc[1]],
               c='k', marker='D', s=25, zorder=10)

    return ax

def plot_commits_by_engineer(self):
        plt.figure(figsize=(16, 14))
        sns.stripplot(x="CommitDate", y="Engineer", data=self.tickets,
                      jitter=True)

def response_surface_analysis(df):
    """Perform response surface analysis on df."""

    def tally_results(df):
        features = [f for f in list(df.columns.values)
                    if f not in ['classifier', 'test_accuracy']]
        classifiers = ['Decision Tree', 'Linear SVC', 'Logistic Regression',
                       'Multinomial NB', 'Random Forest', 'Voting Classifier']
        for classifier in classifiers:
            from_class = df['classifier'] == classifier
            class_mean = df[from_class]['test_accuracy'].mean()
            for feature in features:
                with_feature = df[(df[feature] == 1) & from_class]
                wo_feature = df[(df[feature] == 0) & from_class]
                acc_diff = (with_feature['test_accuracy'].sum()
                            - wo_feature['test_accuracy'].sum())
                mean_diff = acc_diff / 255 + class_mean
                yield classifier, feature, mean_diff

    results = pd.DataFrame([res for res in tally_results(df)],
                           columns=['classifier', 'feature', 'effect'])
    class_order = list(df.groupby('classifier')['test_accuracy']
                         .mean()
                         .sort_values(ascending=False)
                         .index)
    results['classifier'] = pd.Categorical(results['classifier'],
                                           categories=class_order)
    feat_order = list(results.groupby('feature')['effect']
                             .mean()
                             .sort_values(ascending=False)
                             .index)
    results['feature'] = pd.Categorical(results['feature'],
                                        categories=feat_order)
    results.sort_values(['feature', 'classifier'], inplace=True)
    print(results)
    sns.stripplot('effect', 'feature', hue='classifier', data=results)
    sns.plt.legend()
    sns.plt.show()

def aga_attachedness(
        adata,
        attachedness_type='scaled',
        color_map=None,
        show=None,
        save=None):
    """Attachedness of aga groups.
    """
    if attachedness_type == 'scaled':
        attachedness = adata.uns['aga_attachedness']
    elif attachedness_type == 'distance':
        attachedness = adata.uns['aga_distances']
    elif attachedness_type == 'absolute':
        attachedness = adata.uns['aga_attachedness_absolute']
    else:
        raise ValueError('Unkown attachedness_type {}.'.format(attachedness_type))
    adjacency = adata.uns['aga_adjacency']
    matrix(attachedness, color_map=color_map, show=False)
    for i in range(adjacency.shape[0]):
        neighbors = adjacency[i].nonzero()[1]
        pl.scatter([i for j in neighbors], neighbors, color='green')
    utils.savefig_or_show('aga_attachedness', show=show, save=save)
    # as a stripplot
    if False:
        pl.figure()
        for i, ds in enumerate(attachedness):
            ds = np.log1p(ds)
            x = [i for j, d in enumerate(ds) if i != j]
            y = [d for j, d in enumerate(ds) if i != j]
            pl.scatter(x, y, color='gray')
            neighbors = adjacency[i]
            pl.scatter([i for j in neighbors],
                       ds[neighbors], color='green')
        pl.show()

def stripboxplot(x, y, data, ax=None, significant=None, **kwargs):
    """
    Overlay a stripplot on top of a boxplot.
    """
    ax = sb.boxplot(
        x=x,
        y=y,
        data=data,
        ax=ax,
        fliersize=0,
        **kwargs
    )

    plot = sb.stripplot(
        x=x,
        y=y,
        data=data,
        ax=ax,
        jitter=kwargs.pop("jitter", 0.05),
        color=kwargs.pop("color", "0.3"),
        **kwargs
    )

    if data[y].min() >= 0:
        hide_negative_y_ticks(plot)
    if significant is not None:
        add_significance_indicator(plot=plot, significant=significant)

    return plot

def stemplot(x, y, data, order, ax, palette, marker='o', size=7):
    """
    Wrapper to make one stemplot with colored dashed lines leading
    to colored marker.

    Parameters
    ----------
    x, y : str
        used in call to sns.stripplot() with data
    data : pandas dataframe
        Should have the values that are given for 'order' in the index,
        or a column called 'label' with those values.
    order : list
        order of x values
    ax : Axis object
        axis handle to plot values on
    palette : dict
        {values in x-axis : color mapping value}
    marker : str
        marker value to pass to stripplot
    size : int
        size of marker

    Returns
    -------
    ax
    """

    if 'label' in data:
        data.index = data['label']

    sns.stripplot(x=x, y=y, data=data, order=order, ax=ax, palette=palette,
                  size=size, marker=marker)
    _, stemlines, baseline = ax.stem(data.loc[order, y],
                                     markerfmt=" ", linefmt=":")
    # Remove stemplot baseline
    plt.setp(baseline, visible=False)

    # Change stem colors
    colorslist = [palette[i] for i in order]
    _ = [plt.setp(stemlines[i], 'color', colorslist[i])
            for i in range(len(colorslist))]
    _ = [i.set_alpha(0.75) for i in stemlines]

    return ax

def plot_alphadf(alphasdf, col_order, labeldict, metric='alpha'):
    """
    Plot faceted alpha diversity.

    Parameters
    ----------
    alphasdf : pandas DataFrame
        columns ['study', 'alpha', 'DiseaseState']
    col_order : list
        dataset IDs in the order they should be plotted
    labeldict : dict
        dictionary with {dataset: label}
    mteric : str
        alpha diversity metric, to use in labeling y axis

    Returns
    -------
    fig : Figure
    """
    sns.set_style('white')
    g = sns.FacetGrid(alphasdf, col='study', col_wrap=6,
                      col_order=col_order, sharex=False, sharey=False)
    g = g.map(sns.boxplot, "DiseaseState", "alpha")
    g = g.map(sns.stripplot, "DiseaseState", "alpha", split=True, jitter=True,
              size=5, linewidth=0.6)

    fig = plt.gcf()
    fig.set_size_inches(14.2, 9)

    # Fix y-axis gridlines
    axs = g.axes
    for i in range(len(axs)):
        ax = axs[i]
        yticks = ax.get_yticks()
        # If bottom limit is between 0 and 1 (i.e. not simpson)
        if not (yticks[0] < 1 and yticks[0] > 0):
            ax.set_ylim(floor(yticks[0]), floor(yticks[-1]))
        if yticks[0] < 0:
            ax.set_ylim(0, floor(yticks[-1]))

        yticks = ax.get_yticks()
        if (yticks[0] < 1 and yticks[0] > 0):
            ax.set_yticks(yticks[1::2])
        else:
            ax.set_yticks(yticks[::2])
            # Need some space on the y-axis for p-values
            ax.set_ylim(ax.get_ylim()[0], 1.2*ax.get_ylim()[1])
        # Update title
        oldtitle = ax.get_title()
        newtitle = labeldict[oldtitle.split('=')[1].strip()]
        ax.set_title(newtitle)

        # Update y label
        if i % 6 == 0:
            ax.set_ylabel(metric)

    plt.tight_layout()
    return fig

def plot_ubiq_abun_boxplot(tidy, metric, calculation):
    """
    Plot boxplot where x-axis is 'overall_significance' of genus, and values
    are either ubiquity or abundance in tidy (with the respective metric and
    calculation type)

    Parameters
    ----------
    tidy : pandas dataframe
        has columns overall_significance, value, patient, metric, and calculation
    metric : str
        'abundance' or 'ubiquity'
    calculation: str
        'from_pooled_mean' or 'mean_of_datasets'

    Returns
    -------
    ax : Axis object
    """
    fig, ax = plt.subplots(figsize=(5.5,4))
    tmp = tidy.query('metric == @metric')\
              .query('calculation == @calculation')\
              .query('patient == "total"')

    boxprops = {'edgecolor': 'k', 'facecolor': 'w'}
    lineprops = {'color': 'k'}

    # Plot log10(abundance)
    if metric == 'abundance':
        tmp.loc[tmp.index, 'value'] = tmp['value'].apply(np.log10)

    sns.boxplot(data=tmp, x='overall_significance', y='value',
                fliersize=0, ax=ax, color='w',
                order=['health', 'disease', 'mixed', 'not_sig'],
                **{'boxprops': boxprops, 'medianprops': lineprops,
                   'whiskerprops': lineprops, 'capprops': lineprops})

    sns.stripplot(data=tmp, x='overall_significance', y='value',
                  jitter=True, linewidth=0.6, split=True, ax=ax,
                  order=['health', 'disease', 'mixed', 'not_sig'],
                  color='w')
    return fig, ax