Skip to content

roadstyle

roadstyle turns a GeoDataFrame of road edges into a beautifully styled interactive map — an OSM-style "geometry sandwich" (coloured fills over casings) with palettes, settings, and base maps — and can colour roads by your own data, on folium / lonboard / a self-contained MapLibre (vector) backend, or as a stack-agnostic JSON spec you embed in any website.

import geopandas as gpd
from roadstyle import render_edges

edges = gpd.read_file("edges.gpkg")            # any CRS; needs a road-class column
render_edges(edges, basemap="dark_matter").save("roads.html")          # classic OSM styling

render_edges(edges, color_by="aadt", cmap="viridis",          # colour by a data value
             width_by=(1, 6), legend=True).save("traffic.html")

…or straight from the shell — no Python needed:

roadstyle edges.gpkg -o roads.html --basemap dark_matter
roadstyle edges.gpkg --color-by aadt --cmap viridis --width-by 1 6   # colour by your data
roadstyle studio                                                    # or click through it — the Streamlit workbench

Why roadstyle?

  • Geometry sandwich — every road is a coloured fill over a wider casing, so junctions and overlaps read cleanly (just like the OSM "Standard" style).
  • Class styling — built-in highsat (high-saturation) and carto (classic OSM) palettes; bring your own vocabulary with register_palette / palette JSON.
  • Data-driven styling — colour/size roads by any categorical column (color_by+colors) or numeric column (color_by+cmap+width_by), with automatic legends.
  • Base maps — voyager / positron / dark matter / OSM / satellite, plus the tile-less blank canvas (zero network requests — fully offline), with an in-map switcher.
  • 3D view — view_3d=True: tilted camera, extruded ramped bridge decks with black casing, an on-map 2D/3D toggle.
  • Scriptable — every control has a window.rs* JavaScript twin, plus id-set queries (rsQuery → filter / colour / highlight / table / fly-to) and rs:* events; copyable UI templates live in ui/.
  • One settings file — every styling default in data/defaults.json, overridable via roadstyle.json, rs.use_settings(...), or a per-call settings=.
  • Backends + web output — folium (portable HTML), lonboard (WebGL), a self-contained MapLibre web backend (per-zoom widths, two-way lanes, arrows/names, hover/select, tunnel/bridge grade separation — offline, no server; see web backend), and to_spec/to_html/to_iframe for embedding in your own site (Leaflet / MapLibre / iframe).
  • Switchable colouring — bake several "colour by" options with color_options and switch between them in the browser (a Colour by dropdown), no re-render; blank edges keep the base colour.
  • Overlay layers — bring your own geometry (zones, POIs, any lines) as Overlays, drawn under/over the roads, clickable, with a Layers toggle.
  • Canonical input — normalize_edges reprojects, drops non-lines, and maps your column names.
  • Command line — the roadstyle CLI renders any road file from the shell, no Python required.

Where to next

  • Gallery — one screenshot + recipe per look.
  • Manual — a hands-on walk-through with the live map embedded after each step.
  • Usage — install, the command line, quick start, recipes.
  • Choosing an engine — web vs folium vs lonboard vs roadstyle.js, with a feature matrix.
  • Parameter reference — every parameter explained.
  • MapLibre web backend — the self-contained, zoom-correct vector map (two-way lanes, grade separation, offline).
  • Embedding in a website — the JSON spec + Leaflet/MapLibre/iframe snippets.
  • Palettes — styling + settings reference.
  • Comparison — roadstyle vs geopandas .explore() / prettymaps.