Compact layout and Dashboard layout needed

[MarcSkovMadsen]

Problem

The default layout takes up a lot of screen space. So it's difficult to make an app that can be contained in a window, i.e. that does not need scrolling. And a dashboard with multiple widgets and plots is impossible.

Solution

Some kind of "compact" theme or better opportunities to control the layout including some kind og grid layout and opportunities to size the selectboxes, sliders etc.

Additional context

The below python code takes up the full height of my laptop screen. So that's not very good. But imagine I wanted to create a dashboard with multiple chart and widgets.

import pandas as pd
import streamlit as st
import plotly.express as px

data = {"x": [1, 2, 3, 4], "y": [2, 4, 6, 8]}

st.selectbox("Select x", options=["A", "B"])
st.selectbox("Select y", options=["A", "B"])
dataframe = pd.DataFrame(data)
fig = px.scatter(dataframe, x="x", y="x", title="Plot 1", height=400)
st.plotly_chart(fig, height=400)
st.slider("Select min and max", min_value=1, max_value=10, value=[1, 4])

[MarcSkovMadsen]

I know one solution here is to put the widgets in the sidebar. But you don't always wan't to do that.

[asw-v4]

One way to circumvent this could be to create a st.selectbox that you can use to define which visualisation you want to view.

I too would very much appreciate this enhancement, as even this fix can spiral out of control when dealing with massive visualisations

[Boadzie]

I agree with the gentlemen. I will love it if we can have layouts like R flexdashboard. It will be awesome!! Anyway, Thanks guys for a great job. I love Streamlit

[MarcSkovMadsen]

Hi @tvst (and everybody else)

Inspiration for what the basic requirements for layout in streamlit could be can be found here https://ipywidgets.readthedocs.io/en/latest/examples/Layout%20Templates.html.

The AppLayout and GridLayout is really the basics of what we need (I think).

The AppLayout could maybe be implemented something like

with st.header():
    st.write("Awesome Analytics Apps :-)")

with st.right(sidebar=True, width=200):
    st.write("this is my right sidebar. Look just like the existing left st.sidebar")

with st.center():
     st.markdown("# Section 1")

with st.footer():
    st.write("This is my footer").

and the GridLayout something like

with st.grid(columns=2, rows=3):
    with st.cell(col=(1,2), row=1):
        st.write("this spans the first row")
    with st.cell(col=1, row=(2,3)):
        st.write("this spans col 1 of row 2+3")
    with st.cell(col=2, row=(2,3)):
        st.button("this button spans col 2 of row 2+3")

Any comments suggestions are very much appreciated.

[universewill]

mark

[mqyousif]

Indeed would love to be able to position charts and widgets on different parts of the screen, something similar to splitting the screen into different sections PowerBi.

[Boadzie]

Yes! something like that will be great. I love Streamlit and with layout, we have a the perfect framework for Data Science and more

…

On Fri, May 29, 2020 at 4:42 PM mqyousif ***@***.***> wrote: Indeed would love to be able to position charts and widgets on different parts of the screen, something similar to splitting the screen into different sections PowerBi. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#486 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHYPWINSF44BIDS6YQPMWZTRT7QYJANCNFSM4JDC5EVQ> .

[randyzwitch]

Closing, as this is being handled by the horizontal layout functionality that has been released.