mirror of
https://github.com/apache/superset.git
synced 2024-09-12 08:39:45 -04:00
parent
dab31eb66d
commit
fe045ad076
@ -19,6 +19,7 @@ var sourceMap = {
|
||||
markup: 'markup.js',
|
||||
para: 'parallel_coordinates.js',
|
||||
pie: 'nvd3_vis.js',
|
||||
box_plot: 'nvd3_vis.js',
|
||||
pivot_table: 'pivot_table.js',
|
||||
sankey: 'sankey.js',
|
||||
sunburst: 'sunburst.js',
|
||||
@ -45,6 +46,7 @@ var color = function () {
|
||||
// Color factory
|
||||
var seen = {};
|
||||
return function (s) {
|
||||
if (!s) { return; }
|
||||
// next line is for caravel series that should have the same color
|
||||
s = s.replace('---', '');
|
||||
if (seen[s] === undefined) {
|
||||
|
@ -120,6 +120,14 @@ function nvd3Vis(slice) {
|
||||
.staggerLabels(true);
|
||||
break;
|
||||
|
||||
case 'box_plot':
|
||||
colorKey = 'label';
|
||||
chart = nv.models.boxPlotChart();
|
||||
chart.x(function (d) { return d.label; });
|
||||
chart.staggerLabels(true);
|
||||
chart.maxBoxWidth(75); // prevent boxes from being incredibly wide
|
||||
break;
|
||||
|
||||
default:
|
||||
throw new Error("Unrecognized visualization for nvd3" + viz_type);
|
||||
}
|
||||
|
@ -263,6 +263,18 @@ def load_world_bank_health_n_pop():
|
||||
until="now",
|
||||
viz_type='area',
|
||||
groupby=["region"],)),
|
||||
Slice(
|
||||
slice_name="Box plot",
|
||||
viz_type='box_plot',
|
||||
datasource_type='table',
|
||||
table=tbl,
|
||||
params=get_slice_json(
|
||||
defaults,
|
||||
since="1960-01-01",
|
||||
until="now",
|
||||
whisker_options="Tukey",
|
||||
viz_type='box_plot',
|
||||
groupby=["region"],)),
|
||||
]
|
||||
for slc in slices:
|
||||
merge_slice(slc)
|
||||
|
@ -315,6 +315,17 @@ class FormFactory(object):
|
||||
'100',
|
||||
])
|
||||
),
|
||||
'whisker_options': FreeFormSelectField(
|
||||
'Whisker/outlier options', default="Tukey",
|
||||
description=(
|
||||
"Determines how whiskers and outliers are calculated."),
|
||||
choices=self.choicify([
|
||||
'Tukey',
|
||||
'Min/max (no outliers)',
|
||||
'2/98 percentiles',
|
||||
'9/91 percentiles',
|
||||
])
|
||||
),
|
||||
'row_limit':
|
||||
FreeFormSelectField(
|
||||
'Row limit',
|
||||
|
109
caravel/viz.py
109
caravel/viz.py
@ -16,6 +16,7 @@ from collections import OrderedDict, defaultdict
|
||||
from datetime import datetime, timedelta
|
||||
|
||||
import pandas as pd
|
||||
import numpy as np
|
||||
from flask import flash, request, Markup
|
||||
from markdown import markdown
|
||||
from pandas.io.json import dumps
|
||||
@ -488,6 +489,113 @@ class NVD3Viz(BaseViz):
|
||||
is_timeseries = False
|
||||
|
||||
|
||||
class BoxPlotViz(NVD3Viz):
|
||||
|
||||
"""Box plot viz from ND3"""
|
||||
|
||||
viz_type = "box_plot"
|
||||
verbose_name = "Box Plot"
|
||||
sort_series = False
|
||||
is_timeseries = True
|
||||
fieldsets = ({
|
||||
'label': None,
|
||||
'fields': (
|
||||
'metrics',
|
||||
'groupby', 'limit',
|
||||
),
|
||||
}, {
|
||||
'label': 'Chart Options',
|
||||
'fields': (
|
||||
'whisker_options',
|
||||
)
|
||||
},)
|
||||
|
||||
def get_df(self, query_obj=None):
|
||||
form_data = self.form_data
|
||||
df = super(BoxPlotViz, self).get_df(query_obj)
|
||||
|
||||
df = df.fillna(0)
|
||||
|
||||
# conform to NVD3 names
|
||||
def Q1(series): # need to be named functions - can't use lambdas
|
||||
return np.percentile(series, 25)
|
||||
|
||||
def Q3(series):
|
||||
return np.percentile(series, 75)
|
||||
|
||||
whisker_type = form_data.get('whisker_options')
|
||||
if whisker_type == "Tukey":
|
||||
|
||||
def whisker_high(series):
|
||||
upper_outer_lim = Q3(series) + 1.5 * (Q3(series) - Q1(series))
|
||||
series = series[series <= upper_outer_lim]
|
||||
return series[np.abs(series - upper_outer_lim).argmin()]
|
||||
|
||||
def whisker_low(series):
|
||||
lower_outer_lim = Q1(series) - 1.5 * (Q3(series) - Q1(series))
|
||||
# find the closest value above the lower outer limit
|
||||
series = series[series >= lower_outer_lim]
|
||||
return series[np.abs(series - lower_outer_lim).argmin()]
|
||||
|
||||
elif whisker_type == "Min/max (no outliers)":
|
||||
|
||||
def whisker_high(series):
|
||||
return series.max()
|
||||
|
||||
def whisker_low(series):
|
||||
return series.min()
|
||||
|
||||
elif " percentiles" in whisker_type:
|
||||
low, high = whisker_type.replace(" percentiles", "").split("/")
|
||||
|
||||
def whisker_high(series):
|
||||
return np.percentile(series, int(high))
|
||||
|
||||
def whisker_low(series):
|
||||
return np.percentile(series, int(low))
|
||||
|
||||
else:
|
||||
raise ValueError("Unknown whisker type: {}".format(whisker_type))
|
||||
|
||||
def outliers(series):
|
||||
above = series[series > whisker_high(series)]
|
||||
below = series[series < whisker_low(series)]
|
||||
# pandas sometimes doesn't like getting lists back here
|
||||
return set(above.tolist() + below.tolist())
|
||||
|
||||
aggregate = [Q1, np.median, Q3, whisker_high, whisker_low, outliers]
|
||||
df = df.groupby(form_data.get('groupby')).agg(aggregate)
|
||||
return df
|
||||
|
||||
def to_series(self, df, classed='', title_suffix=''):
|
||||
label_sep = " - "
|
||||
chart_data = []
|
||||
for index_value, row in zip(df.index, df.to_dict(orient="records")):
|
||||
if isinstance(index_value, tuple):
|
||||
index_value = label_sep.join(index_value)
|
||||
boxes = defaultdict(dict)
|
||||
for (label, key), value in row.items():
|
||||
if key == "median":
|
||||
key = "Q2"
|
||||
boxes[label][key] = value
|
||||
for label, box in boxes.items():
|
||||
if len(self.form_data.get("metrics")) > 1:
|
||||
# need to render data labels with metrics
|
||||
chart_label = label_sep.join([index_value, label])
|
||||
else:
|
||||
chart_label = index_value
|
||||
chart_data.append({
|
||||
"label": chart_label,
|
||||
"values": box,
|
||||
})
|
||||
return chart_data
|
||||
|
||||
def get_data(self):
|
||||
df = self.get_df()
|
||||
chart_data = self.to_series(df)
|
||||
return chart_data
|
||||
|
||||
|
||||
class BubbleViz(NVD3Viz):
|
||||
|
||||
"""Based on the NVD3 bubble chart"""
|
||||
@ -1387,6 +1495,7 @@ viz_types_list = [
|
||||
IFrameViz,
|
||||
ParallelCoordinatesViz,
|
||||
HeatmapViz,
|
||||
BoxPlotViz,
|
||||
]
|
||||
|
||||
viz_types = OrderedDict([(v.viz_type, v) for v in viz_types_list])
|
||||
|
Loading…
Reference in New Issue
Block a user