metrosp explorer

An interactive dashboard for exploring passenger demand on the São Paulo metro, built with Shiny on top of the metrosp R data package. The interface is in Portuguese.

Try the live app — deployed on Posit Connect Cloud.

The app is standalone: it lives in its own repository and deploys separately from the data package.

Features

• Line-level demand — monthly entrance/transported series per line, with KPIs and an optional STL trend overlay.
• Per-station series — monthly weekday transported averages and daily entries, kept as distinct measures.
• Interactive map — metric views for yearly demand with an animated year slider, year-over-year change, recovery vs. 2019, and the network by line. Station popups show KPIs and link to each station's series.
• Dataset downloads — the package datasets verbatim, in CSV / Excel / GPKG / GeoJSON.

STL uses the full available history within each uninterrupted monthly segment, then applies the selected display period. Missing months and values remain missing; segments with at most 24 observations have no trend estimate.

Run locally

Start the app from the repository root.

shiny::runApp(".")

Dependencies

renv manages the dependencies. The lockfile (renv.lock) pins every package to an exact version and source, and it pins metrosp 2.0.0 to GitHub commit 1bf8859, which fixes archived station-name lookup when sf is loaded.

After cloning, restore the project library.

renv::restore()

Deploy

The repository root is the app, in Shiny's multi-file layout — global.R (libraries, data prep, helpers, theme, sourced once at startup), ui.R, server.R, www/, and a committed manifest.json. It deploys as a unit.

Posit Connect Cloud (git-backed)

Connect Cloud publishes straight from this public GitHub repo. When it detects renv.lock, it calls renv::restore() to install the packages, including the pinned GitHub revision of metrosp. manifest.json stays tracked for app-type metadata, and no rsconnect push is required.

Classic Posit Connect / shinyapps.io

rsconnect::deployApp(appName = "metrosp-explorer")

Updating the lockfile

After adding or upgrading packages, re-snapshot and commit the lockfile.

renv::snapshot()

Data source

The app reads the four rolling demand tables through metrosp::read_metro_demand(). If any rolling table fails, it uses the four bundled snapshots together and identifies that source in the interface. Line and station geometries come from metrosp::rail_lines and metrosp::rail_stations.

The metrosp package documents the datasets and original producers in its data dictionary. The STL trend overlay uses trendseries.

License

MIT © Vinicius Oike