Python_Function for customdata hover in plotly lib

In the code below I tried to use customdata to make hovertemplate, but in this case on visualization it shows only data from the first row everywhere. I believe there should be function, but don't know how to implement it.

import plotly.express as px
import plotly.graph_objs as go
import pandas as pd

rows=[['501-600','15','122.58333','45.36667'],
      ['till 500','4','12.5','27.5'],
      ['more 1001','41','-115.53333','38.08'],
      ]

colmns=['bins','data','longitude','latitude']
df=pd.DataFrame(data=rows, columns=colmns)
df = df.astype({"data": int})

fig=px.scatter_geo(df,lon='longitude', lat='latitude',
                      color='bins',
                      opacity=0.5,
                      size='data',
                      projection="natural earth")

new_customdata = df.loc[:,('bins', 'data')]
fig.update_traces(go.Scattergeo(
customdata=new_customdata,
hovertemplate="<b>%{customdata[0]} </b><br><br>" + \
              "blablabla: %{customdata[1]: .3f}<extra></extra>"))

fig.show()

I believe this does what you're looking for:

rows=[['501-600','15','122.58333','45.36667'],
      ['till 500','4','12.5','27.5'],
      ['more 1001','41','-115.53333','38.08'],
      ]

colmns=['bins','data','longitude','latitude']
df=pd.DataFrame(data=rows, columns=colmns)
df = df.astype({"data": int})

fig = go.Figure(data=go.Scattergeo(
    lon = df['longitude'],
    lat = df['latitude'],
    mode = 'markers', 
    marker_color = df.index, 
    marker_size=df['data'], 
    customdata = df, 

    hovertemplate="<b>%{customdata[0]} </b><br><br>blablabla: %{customdata[1]: .3f}<extra></extra>"

))

fig.show()

The result is (there's different hover text for each item):