Is it possible to sort values within stacked bar graphs?

đź“Š Plotly Python
1

newplot(9)
newplot(9)1904Ă—939 38.7 KB

I want to show the highest value in each bar column to be at the top/end, so for the first one “whale, beluga, meat, dried (Alaska Native)” i need Iron(Cyan color) to be at the very end, etc…

2

Yes there is. You will need to add each row as an individual trace and show the legend only the first time you add a specific category/ color. Use legendgroup to ensure that elements of the same category get selected/ unselected together.

Capture
Capture2495Ă—1105 89.3 KB 

import plotly.express as px
import plotly.graph_objects as go

df = px.data.gapminder()
df = df.loc[df["country"].isin(
      [
         'Germany',
         'United Kingdom',
         'Finland',
         'Belgium',
         'Sweden',
         'Denmark',
         'Austria',
         'Iceland',
         'Netherlands',
         'Switzerland',
         'Ireland',
         'Norway'
      ]
   )
]
df["country_index"] = df.groupby("year")["country"].cumcount()
df = df.sort_values(["year", "gdpPercap"], ignore_index=True)
df["showlegend"] = df.groupby("country")["year"].cumcount() == 0
colors = px.colors.qualitative.Alphabet

fig = go.Figure()
for index, row in df.iterrows():
   fig.add_trace(
      go.Bar(
         y=[row["year"]],
         x=[row["pop"]],
         customdata=[row["gdpPercap"]],
         hovertemplate="%{x:2f}, GDP per capita: %{customdata:2f}",
         marker=dict(color=colors[row["country_index"]]),
         name=row["country"],
         legendgroup=row["country"],
         showlegend=row["showlegend"],
         orientation="h"
      )
   )

fig.update_layout(barmode="stack", hovermode="y unified")