Creating a Stacked Bar Chart with Bokeh

To create a stacked bar chart with Bokeh using your queryset data, you need to transform your data into a format that Bokeh can use for stacked bars. The issue you're encountering is likely because you're trying to use the raw queryset data directly, which has duplicate Order_Type values.

Here's a solution to create the stacked bar chart:

PYTHONCopied
from bokeh.plotting import figure, show
from bokeh.models import ColumnDataSource
from bokeh.palettes import Category10
from collections import defaultdict

# Assuming your queryset data is in a variable called 'queryset'
queryset = [
    {'Order_Type': 'Phone', 'Order_Priority': 'NA', 'Order_total': 2},
    {'Order_Type': 'Web', 'Order_Priority': 'High', 'Order_total': 3},
    {'Order_Type': 'Web', 'Order_Priority': 'Med', 'Order_total': 9}
]

# Group data by Order_Type and Order_Priority
data = defaultdict(dict)
order_types = set()
priorities = set()

for item in queryset:
    order_type = item['Order_Type']
    priority = item['Order_Priority']
    total = item['Order_total']
    
    order_types.add(order_type)
    priorities.add(priority)
    
    data[order_type][priority] = total

# Convert to format needed for stacked bars
order_types = list(order_types)
priorities = list(priorities)

# Create a dictionary for ColumnDataSource
source_data = {'Order_Type': order_types}

# Add data for each priority
for priority in priorities:
    source_data[priority] = [data[order_type].get(priority, 0) for order_type in order_types]

# Create ColumnDataSource
source = ColumnDataSource(source_data)

# Create figure
p = figure(x_range=order_types, height=350, title="Orders by Type and Priority",
           toolbar_location=None, tools="")

# Create stacked bars
colors = Category10[max(3, len(priorities))]
renderers = []

bottom = [0] * len(order_types)
for i, priority in enumerate(priorities):
    r = p.vbar(x='Order_Type', top=priority, width=0.9, source=source,
               bottom=bottom, color=colors[i], legend_label=priority)
    renderers.append(r)
    
    # Update bottom for next stack
    bottom = [bottom[j] + source_data[priority][j] for j in range(len(order_types))]

# Customize figure
p.y_range.start = 0
p.x_range.range_padding = 0.1
p.xgrid.grid_line_color = None
p.axis.minor_tick_line_color = None
p.outline_line_color = None
p.legend.location = "top_right"
p.legend.orientation = "vertical"

# Show the plot
show(p)

This code:

1. Groups your data by Order_Type and Order_Priority
2. Creates a proper data structure for stacked bars
3. Uses ColumnDataSource to feed data to Bokeh
4. Creates the stacked bars with appropriate colors
5. Handles the stacking by keeping track of the "bottom" position for each bar segment

The key to solving your issue is properly transforming the data so that each Order_Type appears only once, with the different Order_Priority values stacked on top of each other.