MapBrief™

If the global economy were collapsing, how would we know?

In mid-February as parts of Asia wrestled with COVID-19, Twitter user @IAMIRONMAN7 had some data to share and he wasn’t shy about connecting the dots:

As I doom-scrolled through the responses, folks who were familiar with Singapore chimed in that this looked like business-as-usual to them, with anecdata reaching High Art in this flex by a Ben Eliott-Yates:

But what troubled me as a geospatial professional is not one of the tweeters in the thread thought to say “there’s probably data that could answer this question.”

And that’s our failure–the failure of the geospatial industry’s marketing message to resonate even faintly with a crowd willing to debate on social media whether Singapore Harbor was crowded or not.  Because our industry’s default setting is to sell rasters and vectors and point clouds and the complicated tools that are supposed to make sense of it all.

Answers?  That’s the customer’s problem.

Bill Emison recounted how too many LiDAR projects concluded with handing the client a hard drive with a few terabytes of LAS files along with the Net-30 invoice.  For most of us the project is over when we zip our Geotiffs or shapefiles, maybe with some metadata, and email the client a download link.

Will Cadell put it more succinctly at last week’s virtual FOSS4G-UK gathering:

Geospatial people excel at building geospatial things for other geospatial people.

While no one ever starved selling pixels to the NGA, the complacent delivery of the same-old/same-old to the legacy set of tried-and-true set of customers (often, a government entity) has limited our market reach as an industry and left the “vision thing” to others.  You can point to new-ish players like Descartes Labs and Orbital Insight who get this and are building businesses around “answers” and not raw data:  I say good on them and that they’re the exceptions that prove the rule.

 *  *  *  *  *  *

So, was Singapore Harbor unusually crowded in mid-February?

As I wondered about that myself, I came across this web map cooked up by Kristof Van Tricht based on Google Earth Engine which houses a bunch of different datasets including the Sentinel SAR stuff that I knew could answer my question.  Let’s go to the animated GIF:

Maybe it’s the hindsight talking, but to this untrained eye the collapse of the global economy doesn’t seem obvious.

The geospatial industry never tires of trumpeting the ubiquity and centrality of Location.  But until we learn how to deliver meaning and not just raw data pushed through a pipe, the potential of the new markets we’ve been telling each other about for years will remain unrealized, likely to be seized by others.

 — Brian Timoney 

Confession:  my gaming career ended in 1979 when Asteroids was released.  Overwhelmed by multi-axis movement and the hyperspace button, I allocated my arcade quarters to analog pinball and turned my back on console gaming.

As my mapping career marches into its third decade, I impatiently await the re-energizing of online mapping that can transform the massive amounts of new geographic data and visualization techniques into something even fractionally compelling as fake gaming landscapes.

In 2013 my frustration with the stale GIS paradigm being stuffed into a browser and called a “portal” was extreme enough to write five blog posts on the topic.  Fast forward to the present, and every couple of months a well-meaning soul will reference the series and remark “and nothing has changed!”

If only I could weaponize an exclamation mark.

While the mapping tilt towards gaming and gaming-like landscapes and experiences has been underway awhile, whether it be Google Maps and Mapbox offering Unity SDKs or ESRI’s CityEngine offering.  But a couple of recent announcements caught my eye:

• LiDAR import support in Unreal Engine
• Cesium teaming up with Epic Games to integrate Cesium’s 3D globe into Unreal Engine 

Now with frames-per-second being one of the most important metrics to be optimized in gaming, one of the advantages of fake landscapes is they can be created, modified, and, well, faked to optimize on frame rate.  Howard raises the question as to whether these optimizations can’t help but come at the expense of geodetic accuracy when using “real” geographic data:

That said, the second announcement is very interesting insofar as it appears to be a global scale effort to make the “real world” work in the gaming context:  I’ll be cheering Cesium on.

Now if the current status quo of online mapping mostly works for you, you’re asking if all of this is “must have” or merely “nice to have”. During a recent webinar put on by L3 Harris, multiple participants talked about the desire for a more immersive experience, etc.  But I would put it more strongly:  in a world where organizations are spending more money for custom data collection– imagery, LiDAR. IoT, etc.—they will soon demand a consumption experience commensurate with their data purchase.  We can’t make our clients spend more on ingredients but continue to serve them generic too-long-under-the-heat-lamp consumption experiences.

  *   *   *   *

Enough with this weak tea hand-wavy thought leadership bro–I want custom terrain in my video games now.

No problem.  Bro.  I hope you like golf.

Because we here at Mapbrief believe in the spirit of DIY.  In The Golf Club 2019  , there are tools for users to create their own courses, etc.  So one particularly motivated user published a tutorial for using USGS LIDAR data and OpenStreetMap data for creating your own real courses.

So let’s celebrate that in some small measure that the era of Bring-Your-Own-Data to your favorite video game has already arrived!

— Brian Timoney 

We are a species of gazers, not clickers:

If you make a tooltip or rollover, assume no one will ever see it.

— Archie Tse, NY Times Graphics Dept

When one of the top creators of general audience data visualizations can only lure 10-15% of their audience into clicking on anything, let alone “diving” into the data, we producers of interactive maps need to take notice. It was an excellent post by Dominikus Baur that deepened the conversation by contrasting the time we creators think about features and functionality with the time spent thinking about audience need.

Read it.

For me, it was deja vu all over again. Four years ago I wrote extensively about interactive map portals and their many flaws:

• a visual grammar derived from desktop GIS unfamiliar to most
• too much content that forces users to sort and sift
• End-user experiences driven by often niche internal requirements
• tools that create confusion rather than lead to insight
• we don’t properly value an end-user’s time

So what can we do?

Static Maps

Just because you’re publishing a map to the web, doesn’t mean it has to be a web map.  If a user is only going to spend 10-15 seconds with your map without interacting, why spend two weeks wrestling with your Javascript?  And the great thing is the focus a static map brings–a single view, a single story: don’t bury the lede.  Most of your web mapping platforms have easy to use Static APIs (Google, Mapbox, CARTO, Mapquest, et al) where a map is just a URL away, and you’re desktop GIS can output something web friendly.  And don’t forget that after many, many years SVG is now a first-class citizen of the web, directly viewable in the browser–and you don’t need fancy software to create a sexy map.

Small Multiples

If users aren’t going to interact with a time slider, give them Small Multiples–they’re Tufte approved!

In fact, small multiples may be a better cognitive choice than a time slider because the eye can jump back and forth much more easily to compare patterns.

Animated GIFs

If we in the mapping industry have over-invested in interactive maps relative to user engagement, we most certainly have under-invested in one of the most popular ways to share content on the web.

The animated map GIF.

An update on ‘visualising a lot’: I’ve re-designed, & added cumulative total to the gif, which I think works better: https://t.co/EtrcRMre9N pic.twitter.com/ynalgnVki9

— Alasdair Rae (@undertheraedar) April 5, 2017

As with small multiples, it’s a natural fit for time series data.  But you can also create the mini story map:

Why #China‘s plan to build a new #SilkRoad runs through #Singapore #dataviz #map #geoeconomics #geostrategy https://t.co/79AqbgExj8 pic.twitter.com/eiRk63hcU1

— Fatma E., PhD (@FatmaEr_PhD) April 3, 2017

More Text Box

You know what interaction your users are completely comfortable with?

Text-based search.

Filter first with text search, and make the search box big and obvious.

 Your Map Is The Exception

Sure it is.

The Engineering Department loves it, right?

If internal users are your most important constituency, fine.  But don’t pretend you’re doing your public users any favors with the overly busy interface that of course looks okay to you.  If public engagement is a priority, review the analytics (you’d be surprised how many shops don’t even do this).  If you’re still unconvinced, there are people to help you understand how exactly your map is used.

 *  *  *  *  *

Look, you’re reading this blog because you believe in the unique power of a map.  But to make a map is also to think about its delivery mechanism:  Mercator started out making money from globes; his buddy Ortelius innovated with his atlas.  We 21st century descendants are blessed, and cursed, with a vast audience armed with a fleeting attention span and a thumb with which to engage with, or quickly dismiss, our wares.

— Brian Timoney