Plotly express multiple dropdown menus in python

Question

I am wondering if someone would be able to help me out. I have searched loads of answers on multiple website to get me where I am and I believe I am close to an answer but am struggling with the last bit.

I am trying to make an interactive graph using plotly express

I have a dataframe that looks like this:

	num	label	color	value1	value 2	value3
0	1	A	red	0.4	4	40
1	2	A	blue	0.2	2	20
2	3	A	green	0.3	3	30
3	4	A	red	0.6	6	60
4	5	A	blue	0.7	7	70
5	6	A	green	0.4	4	40
6	7	B	blue	0.2	2	20
7	8	B	green	0.4	4	40
8	9	B	red	0.4	4	40
9	10	B	green	0.2	2	20
10	11	C	red	0.1	1	10
11	12	C	blue	0.3	3	30
12	13	D	red	0.8	8	80
13	14	D	blue	0.4	4	40
14	15	D	green	0.6	6	60
15	16	D	yellow	0.5	5	50

I would like to plot a graph in plotly that has two dropdown menu's, one that controls which value is being plotted and the other that controls which color is being plotted so for example if the dropdowns said: red and value1, I would get a bar graph of only that data.

My current code is as follow and it sort of works but I am getting values appearing on the graph when I change the color settings.

#Import the dataframe
df_test = pd.read_csv('test.csv')

#Plot a figure using plotly express
fig = px.bar(df_test,  y="value1", )

#Create a button list for the value columns
buttonlist = []
for col in df_test.columns[3:]:
      buttonlist.append(
        dict(
            args=['y',[df_test[str(col)]] ],
            label=str(col),
            method='restyle'
        )
      )

#Create a button list for the colors
buttonlist2 = []
for p in df_test['color'].unique():
      buttonlist2.append(
        dict(
            args=['y',[df_test[df_test['color']==p]]],
            label= 'Color ' + str(p),
            method='restyle'))      

#Update the layout to include the dropdown menus, and to show titles etc
fig.update_layout(title="Test",
      yaxis_title="value",
      xaxis_title="Color",
    updatemenus = [
   dict(buttons = buttonlist, showactive = True),
   dict(buttons = buttonlist2, showactive = True, y = 0.9)
   ])

#Show the figure
fig.show()

Any help would be appreciated

Thanks

update your question to include your sample data as markdown instead of an image. print(df_test.to_markdown()) and paste markdown into question — Rob Raymond
– Rob Raymond, Commented Jul 21, 2021 at 14:08
Apologies I didn't know that was possible thanks for the pointer. I believe I have done that now. — MrObvious
– MrObvious, Commented Jul 21, 2021 at 14:16

Rob Raymond · Accepted Answer · 2021-07-22 07:27:31Z

3

updatemenus are static and each dropdown is independent. Hence to provide a facade of being dynamic and linked, build a static structure that does this
below code shows this, but does not rebuild column selector
IMHO, it's much simpler to switch to dash where it becomes very simple to interact with a figure using callbacks to implement your desired behaviour

fig = px.bar(
    df_test,
    y="value1",
)

# Create a button list for the value columns
buttonlist = [
    {
        "args": [
            {
                "y": [df_test.loc[df_test["color"].eq("red"), col]],
                "updatemenus": [
                    {"buttons":[{"label":"Back","method":"relayout","args":[{}]}]},
                    {
                        "buttons": [
                            {
                                "label": f"Color {c}",
                                "method": "restyle",
                                "args": [
                                    {"y": [df_test.loc[df_test["color"].eq(c), col]]}
                                ],
                            }
                            for c in df_test["color"].unique()
                        ],
                        "y": 0.9,
                    }
                ],
            }
        ],
        "label": str(col),
        "method": "relayout",
    }
    for col in df_test.columns[3:]
]

# Update the layout to include the dropdown menus, and to show titles etc
fig.update_layout(
    title="Test",
    yaxis_title="value",
    xaxis_title="Color",
    updatemenus=[
        {"buttons":buttonlist, "showactive":True},
    ],
)

using dash

from jupyter_dash import JupyterDash
import dash_core_components as dcc
import dash_html_components as html
from dash.dependencies import Input, Output, State
import plotly.express as px
import pandas as pd
import io

df_test = pd.read_csv(io.StringIO(""" num label  color  value1  value2  value3
   1     A    red     0.4        4      40
   2     A   blue     0.2        2      20
   3     A  green     0.3        3      30
   4     A    red     0.6        6      60
   5     A   blue     0.7        7      70
   6     A  green     0.4        4      40
   7     B   blue     0.2        2      20
   8     B  green     0.4        4      40
   9     B    red     0.4        4      40
  10     B  green     0.2        2      20
  11     C    red     0.1        1      10
  12     C   blue     0.3        3      30
  13     D    red     0.8        8      80
  14     D   blue     0.4        4      40
  15     D  green     0.6        6      60
  16     D yellow     0.5        5      50"""), sep="\s+")

# Build App
app = JupyterDash(__name__)

app.layout = html.Div(
    [
        dcc.Dropdown(
            id="color",
            options=[{"label": x, "value": x} for x in df_test["color"].unique()],
            value="red",
        ),
        dcc.Dropdown(
            id="value",
            options=[{"label": x, "value": x} for x in df_test.columns[3:]],
            value="value1",
        ),
        dcc.Graph(id="graph1"),
    ]
)


@app.callback(
    Output("graph1", "figure"),
    Input("color", "value"),
    Input("value", "value"),
)
def update_graph1(color, value):
    print(color, value)
    return px.bar(df_test.loc[df_test["color"].eq(color)], y=value)


# Run app and display result inline in the notebook
app.run_server(mode="inline")

edited Jul 22, 2021 at 7:27

answered Jul 21, 2021 at 16:42

Rob Raymond

31.5k3 gold badges19 silver badges34 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

MrObvious Over a year ago

Thank you very much Rob this is really helpful. After I posted the problem I started to realise why is wasn't working and after studying the way the buttonlists are generated I figured out that as you stated that the updatemenus are static and separate meaning they are not talking to each other. My understanding of Dash is that the outputs can not be saved to a HTML file for sharing which is what I need to be able to do.

Rob Raymond Over a year ago

I have included dash solution - as you can see it becomes very simple...

MrObvious Over a year ago

Thanks again it is indeed much easier, however as the output is not easy to share via email (sharing the code it not an option) it would be tricky to use this in my project.

Rob Raymond Over a year ago

it should be quite simple to deploy to azure, GCP or AWS

MrObvious Over a year ago

Yeah I am sure it is but I am not allowed to use those services. Really thanks for your help though it has been brilliant. Cheers

Collectives™ on Stack Overflow

Plotly express multiple dropdown menus in python

1 Answer 1

using dash

5 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

using dash

5 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related