For the open source devotees out there, I've integrated @OpenFreeMapOrg into the #rstats {mapgl} package. Do `maplibre(style = openfreemap_style("liberty"))` for an excellent street map with 3D buildings; "bright" & "positron" are also available. Map with no API keys required!

This is the most exciting time ever to be working in data, and I'm not talking about AI. 3 years ago, I wrote a database-centric guide in my book for analyzing the full 92 million record 1910 Census. Now, with #rstats and @duckdb? Analyze those 92 million rows in seconds.

All 2.78 million Public Land Survey System (PLSS) Sections - Visualized locally on my laptop - Fan running nice and cool with polygons loading instantly. Must have required expensive software / subscriptions to handle this much data, right? Not at all.

A cartographic use-case for mapgl's TurfJS integration: calculate centroids on-the-fly for polygon labels. Shown here: we map Census data; get the centroid with Turf's center of mass function; then use that for a symbol layer to make a custom label.

In the #rstats {mapgl} package, the `move_layer()` function is really useful for setting the visual order of your map layers. Try it out with this interactive app - it uses the {sortable} package to reorder your layers with a drag-and-drop UI element. Code here: gist.github.com/walkerke/190...

In the first Walker Data workshop on the 2020 Census, I taught how to compare single-year age profiles over time - with tips on plotting all states in parallel! I'm deeply discounting past workshops in advance of new content in 2026. Check them out! walkerdata.gumroad.c...

Update to the {mapgl} layers control: you can set layer groups and custom labels for layers. Shown here: mapping Texas Land Survey System areas as both a fill and a line layer, then grouping them together in the layers control. Now on GitHub, so try it out! #rstats #gis

Travel time-aware isochrones can help you make decisions in real estate, retail analytics, supply chains, emergency services, and a whole lot more. In R, it doesn't take much code to get started. Learn how to create & visualize them in my latest post: walker-data.com/post...

The fastest way to start using & exploring the new 2024 1-year ACS data? R with tidycensus + mapgl. With just a few lines of code, you can pull down any of the thousands of variables available in the new data and explore on an interactive MapLibre map.

Released today: the 2024 1-year American Community Survey estimates. This release brings fresh updates to thousands of demographic variables for areas with 65k+ population, as well as new tables on detailed race and HOA/Condo fees.

Take the time to learn the {mirai} package, an async framework for R. There's a learning curve, but it's worth it. Jobs that used to take hours now take minutes in parallel with mirai. And I'm getting never-before-seen performance in my Shiny apps. mirai.r-lib.org/inde...

I don't typically work in Excel, but I'm using it for a task right now. Appreciate you, Claude, for validating my feelings here...

Client-side centroids, convex / concave hulls, and voronoi polygons in your interactive maps... All in #rstats. Learn how: walker-data.com/mapg...

A "glow" effect for mapping points using #rstats and @Mapbox Thought some of you might enjoy trying this out and modifying it on your own data Code here: gist.github.com/walk...#rstats Inspiration from John Nelson / Esri's firefly style

This is exactly what I love about #rstats: A pain point for many in the community (rJava is notoriously difficult to install) Developers finding ways to solve it, in turn helping out with so many downstream projects

With the help of LLMs, I'm seeing a huge market opportunity for #rstats developers right now: productized Shiny apps. R/Shiny developers are typically incredible at the data analysis part of app development.

On the blog: time-aware isochrones for accessibility mapping with #rstats and @Mapbox tools

My best advice to you if you are concerned about the relevance of your skills in the #AI age: Learn the #LLM APIs. When you use APIs, you go beyond "asking ChatGPT" to actually building solutions that fit your workflows. Some successes I've had:

In the mapgl #rstats package, you can create custom patch shapes for your legends with SVG graphics or even sf objects! In this example, we're using one of the isochrones drawn on the map as the patch shape for the legend. Learn how in my latest post: walker-data.com/post...

I'm clearing out many of my past workshop recordings in anticipation of new trainings in the upcoming year. Right now, you can bundle all of my workshops given in 2023 and 2024 (22.5 hours!) for over $2,000 off! Check them out: walkerdata.gumroad.c...

Yesterday, I posted about time-aware isochrones. Today, I'm showing you how to make them yourselves. I'll walk you through how to create isochrones with predicted traffic then map time-of-day changes with a comparison slider -- in a few dozen lines of R code. walker-data.com/posts/time-a...

On the blog: lasso selection and spatial filtering for your Shiny mapping apps walker-data.com/post... #rstats #webapps #llm #gis

Understanding how location access differs by time of day can be critical for emergency services, delivery networks, and your business. With @mapbox.com tools, we can model this with time-aware isochrones, then visualize on a map with a comparison slider.

#Accessibility modelers using #r5r #rstats, check this GUI for playing around with R5 network. If many people find it useful, I would get signal if I should invest any more free time into it. github.com/e-kotov/r5rgui (uses #mapgl by @kylewalker.bsky.social and #r5r by @urbandemog.bsky.social et al)

mapgl for #rstats supports @protomaps.com PMTiles as a data source in both @maplibre.org and @mapbox.com maps. If you need to visualize large datasets but struggle with hosting / infrastructure requirements, PMTiles may be exactly what you need.

Resource of the day: the Welikia Project (@welikiaproject) Explore what the native ecology of New York City was like before it was New York City. Interactive map: www.welikia.org/map-...

Saw a post from an influencer telling followers to "stop using R for anything -- use Python like a normal person" As a 10+ year R & Python user, the irony is that 2025 is the best time EVER to be using R. Here's why:

On the blog: side-by-side synced maps in a single line of code Learn how to do this and use comparison views in your R Shiny apps here: walker-data.com/post...

In the latest version of the @Mapbox Standard Style, you can enable "landmark icons" to add context to your maps and apps. In #rstats, use the mapgl package to try this feature out: mapboxgl(config = list(basemap = list(showLandmarkIcons = TRUE)))

A great way to learn the new Turf.js implementation in the #rstats mapgl package: Try out the Shiny app in the docs that combines the package's draw control with the `turf_filter()` function.

Learned today about @OpenFreeMapOrg - free, high-quality vector tiles for your maps and apps! Use them in #rstats with mapgl by passing the URL to the style param, e.g.: maplibre(style = "tiles.openfreemap.or...#rstats) Styles include "positron", "bright", and "liberty" with 3D!

Educational Attainment in America: a dot-density map of educational attainment for the population age 25+ across the US, built with @Mapbox. I first wrote this in 2017, updated last year - still works great! Explore: personal.tcu.edu/kyl...

Today, I got to experience both the benefits and pitfalls of vibe coding in one morning. Before leaving for work, I had an idea for class: a Shiny app where students could draw what they *think* is the Dallas-Fort Worth metro area then view everyone's responses live.

As an R package developer I like to keep an eye on package usage. While download counts from CRAN are an imperfect metric, they are still helpful for monitoring activity. So I've charted out yearly downloads for three of my packages, with a projection for 2025.

A free three-hour workshop: Analyzing Data from the 2023 American Community Survey in R www.youtube.com/watc... Links to all code + materials are on the YouTube page. #rstats #demographics #census #data

Mapped: natural change and net migration, US counties. 🗺️ Dark green: higher natural change & net migration; 🗺️ Bright green: higher NC, lower NM; 🗺️ Bright purple: Lower NC, higher NM; 🗺️ Light grey: lower NC & lower NM. What stands out to you? #rstats #tidycensus

Well this was fun. @mapbox.com allows you to customize the appearance of their Standard style with lookup table (LUT) images, but I didn't know how to make one. I asked Claude Opus 4.1, and it one-shotted an app to generate them for me.

On the blog: strategies for mapping dense point data in R. Learn how to use: 🌎 Circle clusters 🌎 Dynamic heat maps 🌎 Point jittering for same-location events Read the post: walker-data.com/post... #rstats #gis #datavisualization

In my latest blog post, I showed how to select points of interest on a map with a lasso then send the results to an #LLM to summarize it. This example is simple, but I think there are wide-ranging practical applications here! Check out the post: walker-data.com/posts/lasso-...

A new release of the {mapgl} #rstats package (0.4.0) is now on CRAN with support for spatial analysis with TurfJS! Featuring: 🌎 #GIS operations like buffer, clip, & erase; 🌎 Spatial filtering; 🌎 Voronoi polygons & convex/concave hulls; And more! walker-data.com/mapg...

Updating my slides for the semester; when I find myself thinking "I could use an updated chart for that topic", @OurWorldInData is the first place I look. Iframe embed in my slides, and we are good to go! ourworldindata.org/

I'm only teaching intro Urban Studies these days, but I like to find ways to bring the latest tech into my class. Example: teaching about scales of analysis with a Mapbox (mapgl) map showing counties that "fade away" into Census tracts as you zoom in. Course materials: github.com/walkerke/urb...

Friendly advice: DO NOT ask your LLM to analyze your data. DO ask your LLM to help you generate code so YOU can analyze your data & audit the results. After using LLMs heavily across a wide range of projects, I think this is critical for anyone working with data.

Awesome to see a new 5-star review come in for my latest workshop series on mapgl! If you're interested in checking it out, I opened up a new discount for social followers: walkerdata.gumroad.c... Learn how to make 3D maps, story maps, high-performance Shiny apps, and more!

Travel-time matrices have applications across emergency services, supply chain optimization, site selection, transportation planning, and more. The `mb_matrix()` function in my mapboxapi #rstats package can get you up and running quickly: walker-data.com/mapb...#rstats

There are a lot of options out there for geocoding. I'm personally fond of @Mapbox's solution, which gets you 100k free geocodes per month. So much so that I wrote an #rstats wrapper for it, which you can use in the {mapboxapi} package. Learn how: walker-data.com/mapb...

Last month’s workshop series on modern web mapping with #rstats went great! We covered 3D basemaps, custom map design, fast filtering in Shiny, story maps, and more. The workshop recordings and tutorials are available - with a discount for all of you! walkerdata.gumroad.c...

The same area in Dallas that is heavily devoted to freeways contains the Eiffel Tower, Les Invalides, and the Arc de Triomphe in Paris. Check out MAPMIRROR by @duncrae, a really fun tool to compare the "same area" in two locations by drawing a polygon. mapmirror.dunrae.com/

A free 3-hour workshop: Mapping and Spatial Analysis with US Census Data in R Learn how to: - Get pre-joined spatial US Census data; - Make interactive shaded, circle, and dot-density maps; - Solve the "AK/HI/PR problem" in US mapping; And more!

Voronoi polygons that dynamically re-draw in-browser as you drag markers around your @mapbox.com map. All with R code. This was a fun one to put together, so I'm sharing the code here for you to try: gist.github.com/walkerke/f4d... #rstats #gis #spatialanalysis