How to show individual values from specific condition of dataframe for each bar using plotly python?-CodePudding

I have a several dataframes with user's order information like that:

A = {
    'user': ['a', 'b', 'c', 'd', 'e', 'f'],
    'country': ['UK', 'US', 'US', 'UK', 'UK', 'US'],
    'order': [1, 2, 2, 4, 4, 4]}

B = {
    'user': ['a', 'b', 'c', 'd', 'e', 'f'],
    'country': ['US', 'US', 'UK', 'FR', 'UK', 'US'],
    'order': [2, 2, 3, 3, 3, 3]}

A = pd.DataFrame(A)
B = pd.DataFrame(B)

	user	country	order
0	a	UK	1
1	b	US	2
2	c	US	2
3	d	UK	4
4	e	UK	4
5	f	US	4

	user	country	order
0	a	US	2
1	b	US	2
2	c	UK	3
3	d	FR	3
4	e	UK	3
5	f	US	3

I am trying to plot this data in such a way that I can see number of users vs number of orders. And also show the (unique) countries related to the order group when I navigate to the bar on plot (pass these countries to hovertemplate function?). I can get list of lists with unique countries of each order group (country_list), but I am struggling to pass the correct list of countries separately to each corresponding bar. Like in the highlighted plot (dataframe:B, order group: 3, number of users: 4) I want it to display unique countries: ['US', 'UK', 'FR'] related to that order group . Can anyone help with it?

data_dict = {"A" : A, "B": B}

fig = go.Figure()

for i in data_dict:
    tmp = data_dict[i]
    X = tmp["order"].value_counts().keys()
    Y = tmp["order"].value_counts().to_list()
    country_list = [tmp[tmp['order']==k]['country'].unique().tolist() for k in X]
    fig = fig.add_trace(go.Bar(x = X,
                               y = Y, 
                               hovertemplate='Users: %s' % (country_list),
                               name = i))
    
fig['layout']['yaxis']['title']='Users Count'
fig.update_xaxes(title='Orders')
fig.update_layout(height=500, width=1000)
fig.update_traces(texttemplate='%{y}', textposition='outside')

Like currently it displays the whole list of unique country lists regardless of order group:

But I want it to display only the list of countries specific to that order group:

CodePudding user response：

To get the hover templates as you need, please change the hovertemplate line to this...

hovertemplate="Users: %{text}", text = country_list,

A bit more explanation - the country_list is a list of lists. For hovertemplate, it is expecting a list for each add_trace. The code is expected in the format mentioned above. Adding just (country_list) was causing the issue. Full updated code and working screenshot below.

import plotly.graph_objects as go
A = {
    'user': ['a', 'b', 'c', 'd', 'e', 'f'],
    'country': ['UK', 'US', 'US', 'UK', 'UK', 'US'],
    'order': [1, 2, 2, 4, 4, 4]}
B = {
    'user': ['a', 'b', 'c', 'd', 'e', 'f'],
    'country': ['US', 'US', 'UK', 'FR', 'UK', 'US'],
    'order': [2, 2, 3, 3, 3, 3]}

A = pd.DataFrame(A)
B = pd.DataFrame(B)
data_dict = {"A" : A, "B": B}

fig = go.Figure()

for i in data_dict:
    tmp = data_dict[i]
    X = tmp["order"].value_counts().keys()
    Y = tmp["order"].value_counts().to_list()
    country_list = [tmp[tmp['order']==k]['country'].unique().tolist() for k in X]
    fig = fig.add_trace(go.Bar(x = X, y = Y, 
                               hovertemplate="Users: %{text}", text = country_list,
                               name = i))    
fig['layout']['yaxis']['title']='Users Count'
fig.update_xaxes(title='Orders')
fig.update_layout(height=500, width=1000)
fig.update_traces(texttemplate='%{y}', textposition='outside')

Output screenshots