[Python] Is there a way to overlay scatterplot over grouped boxplots so they aren't offset...

I'm trying to get the scatter plots to lie ontop of their respective boxplots to act as outlier points. Since plotly's graph_object.box doesn't have a method of inputting precalculated outliers, I've been trying to do it this way. I don't want plotly to calculate the outliers for the purposes of the project. Is there any way to accomplish this by moving the scatterplot, or perhaps a feature of go.Box I overlooked that can do this?

import plotly.graph_objects as go

def create_multiple_boxplots(summary_stats_list, labels, types, title="Multiple Boxplots"):
fig = go.Figure()

color_map = {"Rainy": "blue", "Sunny": "green"}

i=0
for stats, label, type_ in zip(summary_stats_list, labels, types):
fig.add_trace(go.Box(
name=type_,
q1=[stats['Q1']],
median=[stats['Median']],
q3=[stats['Q3']],
lowerfence=[stats['Min']],
upperfence=[stats['Max']],
mean=[stats['Mean']],
boxpoints='all' if 'Outliers' in stats else False,
jitter=0.3,
pointpos=-1.8,
marker=dict(color=color_map[type_]), # Assign color based on type
legendgroup=type_,
showlegend=True if i < 2 else False,
x=[label],
# y=stats.get('Outliers', [])
))
# Add outlier points separately

fig.add_trace(go.Scatter(
x=[label] * len(stats['Outliers']),
y=stats['Outliers'],
mode='markers',
marker=dict(color=color_map[type_], size=8, symbol='circle-open'),
name=f"Outliers - {type_}",
legendgroup=type_,
showlegend=False
))
i+=1

fig.update_layout(title=title, yaxis_title="Value", boxmode='group')

fig.show()

# Example summary statistics
data_summaries = [
{"Min": 5, "Q1": 10, "Median": 15, "Q3": 20, "Max": 25, "Mean": 16, "Outliers": [2, 27]},
{"Min": 6, "Q1": 11, "Median": 16, "Q3": 21, "Max": 26, "Mean": 17, "Outliers": [3, 28]},
{"Min": 4, "Q1": 9, "Median": 14, "Q3": 19, "Max": 24, "Mean": 15, "Outliers": [1, 26]},
{"Min": 7, "Q1": 12, "Median": 17, "Q3": 22, "Max": 27, "Mean": 18, "Outliers": [4, 29]}
]
labels = ["Happy", "Happy", "Sad", "Sad"]
types = ["Rainy", "Sunny", "Rainy", "Sunny"]

create_multiple_boxplots(data_summaries, labels, types)





Continue reading...