Aggregate hover data for internal nodes in hierarchical plots

Hierarchical plots constructed with px.treemap, px.icicle, and px.sunburst, aggregate data at internal nodes as follows:

• values: sum of children,
• color: values-weighted sum of children if df[color] is numeric, else "discrete aggregation",
• hover_data: only supports "discrete aggregation",
  where for "discrete aggregation", the value of an internal node is mapped to that of its children if all its children have the same value, otherwise the aggregated value is set to "(?)".

Below is an example of a treemap with 2 quantities in hover data, one of which is numeric and could be aggregated at internal nodes, but instead all non-leaf nodes get a "(?)":

df = px.data.gapminder().query("year == 2007")
df["quantity_to_aggregate"] = np.random.random(df.shape[0])

fig = px.treemap(
    df,
    path=[px.Constant("world"), "continent", "country"],
    values="pop",
    color="lifeExp",
    hover_data=["iso_alpha", "quantity_to_aggregate"],
    color_continuous_scale="RdBu",
    color_continuous_midpoint=np.average(df["lifeExp"], weights=df["pop"]),
)
fig.update_layout(margin=dict(t=50, l=25, r=25, b=25))
fig.show()

Is it possible to aggregate numeric hover_data in hierarchical plots, perhaps as a sum initially, but potentially also giving the option to apply a different function?

Commencez par le comportement des graphiques hiérarchiques décrit pour px.treemap, px.icicle et px.sunburst, en utilisant l’exemple Gapminder fourni avec des hover_data numériques. Déterminez comment hover_data est actuellement agrégé pour les nœuds internes et comment une fonction d’agrégation pourrait être exposée. Le travail est considéré comme terminé lorsque les champs de survol numériques sont agrégés sur les nœuds internes, tandis que les champs discrets conservent leur comportement actuel, et que l’option demandée est prise en charge.