Adding error_y from two columns in a stacked bar graph, plotly express

1
votes

I have created a stacked bar plot using plotly.express. Each X-axis category has two correspondent Y-values that are stacked to give the total value of the two combined.

How can I add an individual error bar for each Y-value? I have tried several options that all yield the same: The same value is added to both stacked bars. The error_y values are found in two separate columns in the dataframe: "st_dev_PHB_%" and "st_dev_PHV_%" , respectively, which correspond to 6 categorical values (x="C").

My intuition tells me its best to merge them into a new column in the dataframe, since I load the dataframe in the bar plot. However, each solution I try give an error or that the same value is added to each pair of Y-values.

What would be nice, is if it's possible to have X error_y values corresponding to the X number of variables loaded in the y=[...,...] . But that would off course be too easy .........................

data_MM = read_csv(....)

#data_MM["error_bar"] = data_MM[['st_dev_PHB_%', 'st_dev_PHV_%']].apply(tuple, axis=1).tolist()

#This one adds the values together instead of adding them to same list.
#data_MM["error_bar"] = data_MM['st_dev_PHB_%'] + data_MM['st_dev_PHV_%']

#data_MM["error_bar"] = data_MM[["st_dev_PHB_%", "st_dev_PHV_%"]].values.tolist()

#data_MM["error_bar"] = list(zip(data_MM['st_dev_PHB_%'],data_MM['st_dev_PHV_%']))

bar_plot = px.bar(data_MM, x="C", y=["PHB_wt%", "PHV_wt%"], hover_data =["PHA_total_wt%"], error_y="error_bar")
bar_plot.show()

The most commonly endured error message:

A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

1
pandasplotlybar-chart

1 Answers

0
votes

I see your problem with the same error bar being used in both bars in the stack. However, I got a working solution with Plotly.graph_objs. The only downside was the second bar is plotting at the front, and therefore the upper half of the lower error bar is covered. At least you can still read off the error value from the hover data.

Here is the full code:

n = 20
x = list(range(1, n + 1))
y1 = np.random.random(n)
y2 = y1 + np.random.random(n)
e1 = y1 * 0.2
e2 = y2 * 0.05

trace1 = go.Bar(x=x, y=y1, error_y=dict(type='data', array=e1), name="Trace 1")
trace2 = go.Bar(x=x, y=y2, error_y=dict(type='data', array=e2), name="Trace 2")

fig = go.Figure(data=[trace1, trace2])
fig.update_layout(title="Test Plot", xaxis_title="X axis", yaxis_title="Y axis", barmode="stack")

fig.show()

Here is a resulting plot (top plot showing one error value, bottom plot showing different error value for the same bar stack): Plot1 Plot2