async def flow_vs_intensity_violin(experiment, intensity=130): data = await flow_vs_intensity_data(experiment, intensity) df_cat_expand = ["Dark"] * len(data["Dark"]) + ["Light"] * len( data["Light"]) df_val_expand = data["Dark"] + data["Light"] df = pd.DataFrame(dict(Flow=df_val_expand, Category=df_cat_expand)) return ff.create_violin(df, data_header="Flow", group_header="Category", title=None, rugplot=False)
async def flow_vs_category_violin(experiment): labels = ["Undefined", "Unknown", "Bitumen", "Sand", "Bubble"] data = [] for category in range(5): data.append(await flow_vs_category_data(experiment, category)) labels = ((["Undefined"] * len(data[0])) + (["Unknown"] * len(data[1])) + (["Bitumen"] * len(data[2])) + (["Sand"] * len(data[3])) + (["Bubble"] * len(data[4]))) df = pd.DataFrame( dict(Flow=data[0] + data[1] + data[2] + data[3] + data[4], Category=labels)) return ff.create_violin(df, data_header="Flow", group_header="Category", title=None, rugplot=False)
def compare_class_clusters_violin(self): cluster_int_labels = np.array(self.cluster_results.labels_) class_values = self.df_class[self.target_key] df_colour_dict = dict([ (class_label, i) for (i, class_label) in enumerate(set(class_values)) ]) class_int_labels = np.array( [df_colour_dict[key] for key in class_values], dtype=int) df = pd.DataFrame() df['Class'] = class_values df['Cluster'] = cluster_int_labels fig = ff.create_violin(df, data_header='Cluster', group_header='Class', height=500, width=800) return py.iplot(fig, filename='Multiple Violins')
def create_violin(*args, **kwargs): FigureFactory._deprecated('create_violin') from plotly.figure_factory import create_violin return create_violin(*args, **kwargs)
df.head() print(df) from plotly.offline import iplot import plotly.figure_factory as ff figure = ff.create_scatterplotmatrix(df[['KnownSex', 'Head', 'Mass', 'Skull']], index='KnownSex', height=800, width=800) iplot(figure) corrs = df.corr() figure = ff.create_annotated_heatmap(z=corrs.round(2).values, x=list(corrs.columns), y=list(corrs.index), showscale=True) #iplot(figure) figure = ff.create_scatterplotmatrix(df[['KnownSex', 'Head', 'Mass', 'Skull']], diag='histogram', index='KnownSex', height=800, width=800) #iplot(figure) iplot(ff.create_violin(df, data_header='Mass', group_header='KnownSex')) figure = ff.create_dendrogram(df[['Mass', 'Skull']]) iplot(figure)
def Plot(self): freqs = self.Frequencies(counts=True, percents=True) df = pandas.DataFrame({"title":[x[0] for x in self.data], "observed stars":[x[1] for x in self.data], "sentiment score":[x[4] for x in self.data]}) group_stats = {} groupby_data = df.groupby(["observed stars"]) for group in [0.0,0.5,1.0,1.5,2.0,2.5,3.0,3.5,4.0]: group_data = groupby_data.get_group(group)["sentiment score"] mdn = numpy.median(group_data) group_stats[group] = mdn box0 = dict(type="box", y=[x[2] for x in self.data], boxmean="sd", name="Measured Stars",marker={"color":"#9d87ed"},xaxis="x1",yaxis="y2") box1 = dict(type="box", y=df["observed stars"], boxmean="sd", name="Observed Stars",marker={"color":"#87aaed"}, xaxis="x1",yaxis="y2") obsbars = dict(type="bar", x=[k for k in freqs["Observed_Perc"].keys()], y=[v for v in freqs["Observed_Perc"].values()], name='Observed Star Ratings', xaxis="x2", yaxis="y1",marker={"color":"#87aaed"}) measbars = dict(type="bar", x=[k for k in freqs["Measured_Perc"].keys()], y=[v for v in freqs["Measured_Perc"].values()], name='Measured Star Ratings', xaxis="x1", yaxis="y1", marker={"color":"#9d87ed"}) diffs = dict(Counter(list(map(lambda x, y: x-y, [float(j[2]) for j in self.data], [float(i[1]) for i in self.data])))) diffbars = graphs.Bar(x=[x for x in diffs.keys()], y=[y for y in diffs.values()], marker=dict(color="#ed8787")) fig = tools.make_subplots(rows=4, cols=2, specs=[[{}, {'rowspan': 2}], [{}, None], [{'rowspan': 2, 'colspan': 2}, None], [None, None]], print_grid=False, subplot_titles=('Measured Star Rating Groups', 'Measured & Observed Stars Box Plot with Median', 'Observed Star Rating Groups', 'Difference Counts (Measured - Observed)')) fig.append_trace(obsbars, 1, 1) fig.append_trace(measbars, 2, 1) fig.append_trace(box0, 1, 2) fig.append_trace(box1, 1, 2) fig.append_trace(diffbars, 3, 1) fig["layout"].update(showlegend=False, title="Observed v Measured Stars") violin = figfactory.create_violin(df, data_header="sentiment score", group_header="observed stars", height=500, width=800, group_stats=group_stats, use_colorscale=True, title="Sentiment Score to Observed Stars") # pltly.plot(fig) # sleep(2) pltly.plot(violin)
def generate_distribution_plot(plot_data, nbins=20, ptype='hist'): dtypes = set() for s_name in plot_data: dtypes.update(plot_data[s_name]) figs = [] for dtype in sorted(dtypes): pdata = [] for s_name in plot_data: try: pdata.append(float(plot_data[s_name][dtype])) except: pass if ptype == 'hist': figs.append( go.Histogram( x = pdata, opacity = 0.75, nbinsx = nbins, name = "{} ({})".format(dtype, len(pdata)) ) ) elif ptype == 'boxplot': figs.append( go.Box( y = pdata, name = "{} ({})".format(dtype, len(pdata)) ) ) elif ptype == 'dotplot': figs.append( go.Box( y = pdata, name = "{} ({})".format(dtype, len(pdata)), boxpoints = 'all', jitter = 0.5, pointpos = 0, line = dict(width = 0), whiskerwidth = 0, fillcolor = 'rgba(0, 0, 0, 0)' ) ) elif ptype == 'violin': if len(figs) == 0: figs = {} dname = "{} ({})".format(dtype, len(pdata)) figs[dname] = pdata else: return 'Error - unrecognised plot type: {}'.format(ptype) layout = {} if ptype == 'hist': layout = go.Layout( barmode = 'overlay', showlegend = True, xaxis = dict( title = "Data" ), yaxis = dict( title = "Sample Count" # TODO - integers only ) ) if ptype == 'violin': figure = ff.create_violin(figs) else: figure = go.Figure(data = figs, layout = layout) plot_div = py.plot( figure, output_type = 'div', show_link = False, config = dict( modeBarButtonsToRemove = [ 'sendDataToCloud', 'resetScale2d', 'hoverClosestCartesian', 'hoverCompareCartesian', 'toggleSpikelines' ], displaylogo = False ) ) return plot_div
print("The mean is %r" % mean) print("The standard deviation is %r" % st_dev) median = np.median(data['alcohol']) maximum = np.max(data['alcohol']) minimum = np.min(data['alcohol']) print("The median is %r" % median) print("The maximum is %r" % maximum) print("The minimum is %r" % minimum) # -------------------------------------------------- y = data['alcohol'].values.tolist() fig = FF.create_violin(y, title='Violin Plot', colors='#604d9e') py.offline.plot(fig, filename='alcohol-violin-visual.html') # -------------------------------------------------- # https://plot.ly/python/box-plots/ y = data['alcohol'].values.tolist() trace = go.Box( y=y, name='Box Plot', boxpoints='all', jitter=0.3, marker=dict(color='rgb(214,12,140)', ), ) layout = go.Layout(
def plot_a_graph(g, secondary_edges=None, title=None, smart_sizing=True, smart_shaping=True, smart_layout=False, mark_root=False, interactive_mode_flag=True, save_image_path=None): if title is None: title = 'StoryGraph<br>' node_color = 'rgba(152, 76, 0, .1)' root_color = 'rgba(0, 152, 0, .15)' if 'pos' not in list(g.nodes(data=True))[0][1].keys(): nx.set_node_attributes( g, name='pos', values=nx.spring_layout( g, iterations=20, # default: 50 k=0.9, # Optimal distance between nodes. If None the distance is set to 1/sqrt(n) )) edge_traces = [] node_violin_traces = [] middle_node_trace = go.Scatter(x=[], y=[], text=[], mode='markers', hoverinfo='text', marker=go.scatter.Marker(opacity=0, ), textposition='top right') edge_weights = [ e[2]['weight'] for e in g.edges(data=True) if e[2]['weight'] > 0 ] min_edge_weight = min(edge_weights) max_edge_weight = max(edge_weights) if max_edge_weight == min_edge_weight: # corner-case min_edge_weight = 0 primary_edge_traces, middle_node_trace = create_edge_trace_for_plotly( g, middle_node_trace=middle_node_trace, min_edge_weight=min_edge_weight, max_edge_weight=max_edge_weight, color='#888', dash=None, smart_layout=smart_layout) edge_traces.extend(primary_edge_traces) if secondary_edges is not None: secondary_edge_traces, middle_node_trace = create_edge_trace_for_plotly( secondary_edges, middle_node_trace=middle_node_trace, min_edge_weight=min_edge_weight, max_edge_weight=max_edge_weight, color='#0ff', dash='dash', smart_layout=smart_layout) edge_traces.extend(secondary_edge_traces) node_trace = go.Scatter(x=[], y=[], text=[], hovertext=[], mode='markers+text', textposition='middle center', textfont={"size": []}, hoverinfo='text', marker=go.scatter.Marker( colorscale='YlGnBu', reversescale=True, color=[], size=[], )) layout = go.Layout( autosize=False, height=(720 if interactive_mode_flag else 960) * np.ceil(len(list(g.nodes())) / 100), width=1320 if interactive_mode_flag else 1760, title=title, titlefont=dict(size=16), font=dict(size=18), # 24 for saving the fig showlegend=False, hovermode='closest', margin=dict(b=20, l=10, r=10, t=60 + 20 * title.count("<br>")), xaxis=go.layout.XAxis(showgrid=False, zeroline=False, showticklabels=False), yaxis=go.layout.YAxis(showgrid=False, zeroline=False, showticklabels=False), shapes=[]) if smart_layout: layout['xaxis'] = go.layout.XAxis(type="date", showgrid=True, ticklen=5) layout['margin'] = dict(b=60, l=5, r=5, t=60 + 20 * title.count("<br>")) # first_trace_flag = True # modified this after switching to plotly 3.3 for node_name, node_data in g.nodes(data=True): x, y = node_data['pos'] # degree = g.degree(node, weight=None) if smart_layout: # x = (node_data['ts_start_epoch'] + node_data['ts_end_epoch']) / 2 * 1000 x = np.mean(node_data['article_epochs']) * 1000 if smart_sizing: if (not smart_shaping) or (len(set(node_data['article_epochs'])) < 3): layout['shapes'] += ({ 'type': 'circle', 'xref': 'x', 'yref': 'y', 'x0': node_data['ts_start_epoch'] * 1000, 'y0': y - max(10, 4 * np.sqrt(node_data['support'])) / 2, 'x1': node_data['ts_end_epoch'] * 1000, 'y1': y + max(10, 4 * np.sqrt(node_data['support'])) / 2, 'fillcolor': root_color if mark_root and node_data['level'] == 0 else node_color, 'line': { 'width': 0.1, }, }, ) else: # https://github.com/plotly/plotly.py/blob/master/plotly/figure_factory/_violin.py if len(node_data['article_epochs']) > 1: node_violin_trace = ff.create_violin( [(article_epoch - node_data['ts_start_epoch']) / (node_data['ts_end_epoch'] - node_data['ts_start_epoch']) for article_epoch in node_data['article_epochs']], rugplot=False)['data'] node_violin_trace = node_violin_trace[: 2] # the first two Scatter objects are the top/bottom halves first_trace_flag = True # dirty trick! modified this after switching to plotly 3.3 for violin_trace in node_violin_trace: temp = np.copy(violin_trace['x']) violin_trace['x'] = 1000 * ( (node_data['ts_end_epoch'] - node_data['ts_start_epoch']) * np.copy(violin_trace['y']) + node_data['ts_start_epoch']) violin_trace['y'] = y + 2 * np.sqrt( node_data['support']) * temp if 'box' in violin_trace: violin_trace['box'] = {'visible': False} if 'scalemode' in violin_trace: violin_trace['scalemode'] = 'count' if 'meanline' in violin_trace: violin_trace['meanline'] = {'visible': False} if 'opacity' in violin_trace: violin_trace['opacity'] = 0.05 if 'fill' in violin_trace: if first_trace_flag: violin_trace['fill'] = 'toself' first_trace_flag = False else: violin_trace['fill'] = 'tonexty' if 'text' in violin_trace: violin_trace['text'] = None if 'line' in violin_trace: violin_trace['line']['color'] = 'rgba(0,0,0,0)' if 'marker' in violin_trace: violin_trace['marker']['color'] = 'rgba(0,0,0,0)' if 'fillcolor' in violin_trace: violin_trace[ 'fillcolor'] = root_color if mark_root and node_data[ 'level'] == 0 else node_color node_violin_traces.extend(node_violin_trace) else: if mark_root and node_data['level'] == 0: node_trace['marker']['color'] += (root_color, ) else: node_trace['marker']['color'] += (node_color, ) # node_data['bet_cent'] if smart_shaping: node_trace['marker']['size'] += (max( 50, 20 * np.sqrt(node_data['support'])), ) # g.node[node]['pr'] else: node_trace['marker']['size'] += (50, ) node_trace['x'] += (x, ) node_trace['y'] += (y, ) node_trace['hovertext'] += ( "<br>".join([ # "%s" % node_name, "%d article%s" % (node_data['support'], "s" if node_data['support'] > 1 else ""), "query relevance: %.4f" % node_data['score'], ("level: %d" % node_data['level']) if 'level' in node_data.keys() else "", "%d day%s" % (node_data['ts_n_dates'], "s" if node_data['ts_n_dates'] > 0 else ""), "%s - %s" % (time.strftime('%d.%m.%Y', time.localtime(node_data['ts_start_epoch'])), time.strftime('%d.%m.%Y', time.localtime(node_data['ts_end_epoch']))), ("redundant tags: %s" % node_data['redundant_nodes']) if 'redundant_nodes' in node_data.keys() else "" # "%s<br>(%.1f %.1f)" % (node_name, x, y) ]), ) node_trace['text'] += ( "%s" % node_name, # "%s<br>(%.1f %.1f)" % (node_name, x, y) ) # + " (degree: %d) (hub_degree: %d)" % (T.degree(node), int(T.node[node]['hub_degree']))) node_trace['textfont']['size'] += (max( 12, 4 * np.log2(node_data['support'])), ) fig = go.Figure(data=[ *node_violin_traces, *edge_traces, node_trace, middle_node_trace ], layout=layout) if interactive_mode_flag: div = plotly.offline.iplot( fig, image_height=720 * np.ceil(len(list(g.nodes())) / 100), image_width=1320, ) else: div = plotly.offline.plot( fig, auto_open=False, output_type='div', image_height=960 * np.ceil(len(list(g.nodes())) / 100), image_width=1760, ) # save the plot(s) if save_image_path is not None: # remove the title fig['layout']['title'] = None fig['layout']['margin'] = dict(b=60, l=40, r=40, t=60) plotly.offline.plot( fig, filename=save_image_path, image='svg', auto_open=False, image_height=480, # 960, image_width=1200, # 1760 ) return div