Shiny for Python: Building Data Apps in Both Languages

Posit's Shiny for Python has reached maturity, bringing the beloved reactive framework to Python developers. Now you can build interactive dashboards using the same powerful paradigm in both R and Python—or even combine them in a single application.

Why Shiny for Python?

Shiny's reactive programming model makes building interactive applications intuitive:

• No JavaScript required for interactivity
• Automatic UI updates when data changes
• Familiar syntax for R Shiny users
• Full Python ecosystem integration

Your First Shiny Python App

from shiny import App, render, ui

app_ui = ui.page_fluid(
    ui.input_slider("n", "Sample size", 10, 1000, 100),
    ui.output_plot("histogram")
)

def server(input, output, session):
    @output
    @render.plot
    def histogram():
        import numpy as np
        import matplotlib.pyplot as plt
        
        data = np.random.randn(input.n())
        plt.hist(data, bins=30, color='#667eea')
        plt.title(f"Histogram of {input.n()} samples")

app = App(app_ui, server)

Comparing R and Python Shiny

R Version

library(shiny)

ui <- fluidPage(
  sliderInput("n", "Sample size", 10, 1000, 100),
  plotOutput("histogram")
)

server <- function(input, output, session) {
  output$histogram <- renderPlot({
    hist(rnorm(input$n), col = "#667eea",
         main = paste("Histogram of", input$n, "samples"))
  })
}

shinyApp(ui, server)

Python Version

The syntax is remarkably similar, making it easy for R users to transition:

# Python uses decorators instead of render functions
@output
@render.plot
def histogram():
    # Python plotting code here
    pass

Advanced Features

Reactive Values

from shiny import reactive

@reactive.Calc
def filtered_data():
    df = load_data()
    return df[df['category'] == input.category()]

@output
@render.table
def data_table():
    return filtered_data()

Modules for Reusability

from shiny import module

@module.ui
def chart_ui(id):
    return ui.card(
        ui.output_plot(id)
    )

@module.server
def chart_server(input, output, session, data):
    @output
    @render.plot
    def plot():
        return create_chart(data())

Mixing R and Python

With Shiny for Python and reticulate, you can use both languages:

# In R Shiny, call Python code
library(reticulate)

output$ml_prediction <- renderText({
  py_model <- import("sklearn.ensemble")
  model <- py_model$RandomForestClassifier()
  # Use Python ML with R Shiny UI
})

Deployment Options

• Posit Connect — Enterprise deployment
• shinyapps.io — Cloud hosting
• Docker — Container deployment
• Hugging Face Spaces — Free hosting for demos

Rflow for Shiny Development

library(rflow)

# Generate Shiny apps from descriptions
rflow_ask("Create a Shiny dashboard with a sidebar 
           for filtering and three chart panels")

# Debug reactive issues
rflow_ask("Why isn't my reactive expression updating 
           when the input changes?")

# Convert between R and Python Shiny
rflow_ask("Convert this R Shiny app to Python Shiny")

Conclusion

Shiny for Python opens new possibilities for data scientists who work in both languages. Whether you're building a quick prototype or a production dashboard, Shiny's reactive model makes interactive data apps accessible to everyone.

Get started: pip install shiny