Chapter 10 Deploying R Shiny Apps on Posit Connect Cloud

If Bookdown is for sharing static documents, Shiny is for sharing interactive tools. A Shiny app lets your users explore data, filter maps, adjust parameters, and see results in real time — all through a web browser, with no R installation required on their end.

For GIS work, Shiny is exceptionally powerful. An interactive map dashboard can communicate spatial patterns far more effectively than a static PDF, and it lets non-technical stakeholders engage with your analysis on their own terms.

10.1 What is Shiny?

• Shiny is an R package for building interactive web applications directly from R — no HTML, CSS, or JavaScript knowledge required (though Claude can add those if you want to customise).
• It’s reactive. When a user changes an input (selects a borough, adjusts a slider, picks a date range), the outputs update automatically.
• It integrates with spatial packages. Leaflet for interactive maps, plotly for interactive charts, DT for searchable data tables — all work seamlessly within Shiny.
• Common GIS use cases:
  • Interactive map dashboards showing spatial indicators by region
  • Data explorers that let users filter and download subsets of spatial data
  • Scenario comparison tools (e.g., “what if we placed a new facility here?”)
  • Monitoring dashboards showing real-time or regularly updated spatial data

10.2 Building a GIS Shiny App with Claude

Claude Code is remarkably effective at building Shiny apps. The combination of UI layout, server logic, and reactive programming is exactly the kind of structured, pattern-based code it handles well.

10.2.1 Getting started

“Create a basic Shiny app that displays a Leaflet map of ward boundaries from data/processed/wards.gpkg, coloured by population density, with a dropdown to select different variables.”

Claude will generate an app.R file (or split ui.R and server.R files) with:

• A sidebar panel with input controls
• A main panel with the Leaflet map
• Server logic that reads the data and renders the map reactively

10.2.2 Adding interactivity

Build up the app’s features incrementally:

• Filtering: “Add a dropdown to filter wards by borough name.”
• Click events: “When the user clicks on a ward, show a popup with the ward name, population, and area.”
• Dynamic legends: “Update the legend when the user selects a different variable.”
• Data tables: “Add a data table below the map showing the attributes of the currently visible wards.”
• Download buttons: “Add a button that lets users download the currently filtered data as a CSV.”

10.2.3 Styling

“Improve the app’s appearance — use a clean layout with a navbar, add a title, and make the map fill the full width of the main panel.”

“Use the bslib package to apply a modern Bootstrap theme.”

10.3 Testing Locally

Before deploying, always test your app locally:

“Run the Shiny app locally so I can test it.”

Claude will run shiny::runApp() and tell you the local URL (usually http://127.0.0.1:XXXX). Check:

• Does the map render correctly?
• Do all inputs and filters work as expected?
• Does it handle edge cases (e.g., filtering to a borough with no data)?
• Is the performance acceptable with your full dataset?

10.4 Deploying to Posit Connect Cloud

The deployment process is similar to Bookdown:

1. Make sure rsconnect is configured (see Chapter 9 if you haven’t done this yet).

2. Deploy the app: > “Deploy this Shiny app to Posit Connect Cloud.”

Claude will run the appropriate rsconnect::deployApp() command with your app directory.

3. Check the deployed version — open the URL Posit Connect gives you and test it in the browser. Deployed apps sometimes behave slightly differently to local ones (file paths, data access, package availability).

10.5 Common Deployment Issues

A few things that catch people out:

• File paths. Your app needs to reference data files using relative paths from the app directory. Absolute paths (~/projects/...) will break on the server.
• Large datasets. Posit Connect Cloud has resource limits. If your app loads a massive spatial dataset on startup, it may time out. Consider pre-processing your data to reduce size, or loading it from a URL.
• Missing packages. Every package your app uses needs to be listed. Claude will usually handle this correctly, but check the deployment logs if something fails.
• Data updates. If your data changes regularly, you’ll need a strategy for refreshing it — either redeploying the app with new data, or having the app read from an external source.

10.6 Updating Your App

As with Bookdown, updating a deployed Shiny app is straightforward:

• Make your changes locally (on a branch!).
• Test locally.
• Redeploy — it updates the same URL.
• Ask Claude: “Redeploy the Shiny app to Posit Connect Cloud.”

The iterative workflow from Chapter 4 applies here too: branch, develop, test, commit, deploy, merge. Your users always see the latest version at the same URL, and you have full version history through Git.