Nicholas Duggan, Dragons8mycat
Thanks for the opportunity, let me first get into smug mode and say that I wasn’t too far off the mark with Boundless’ OpenGeo being prime place for 2014. Even if it was trumped by the presence of CartoDB this year, people were still talking about it. So, CityEngine never really exploded the way I was expecting, it was beaten by ArcGIS Pro and the fantastic QGIS2Threejs which is still a win in my book.
My thought is that 2015 is going to see QGIS become the cartographers’ software of choice, already with many great effects and styling choices. The next year is only going to see those tools expanding and getting more relevant for the cartographer & casual mapper alike.
With the latest Snapdragon chipsets in the current Samsung & HTC devices, locational accuracy can be better than 3 metres anywhere in the world through an off -the-shelf mobile device. I can see this being a springboard for many GIS companies to bring out better mobile mapping/ tracking solutions and also the return of “job manager” systems to quality-control the input & output of this sourced data.
With the UK testing autonomous cars next year, I can see huge advances in LiDAR & data collection appearing… and maybe the reappearance of the indoor mapping…
- 2015 will be the year of specialized personalized OSM maps EVERYWHERE
- maps on e-paper watches
- OSM maps going to Mars. There will be a hashtag #mapstomars
- maps on hats
- maps on those jackets you put on little dogs to keep them from shivering
- maps on free tissues given out in Japanese urban areas
- 2015 will see the first major public lawsuit against the OpenStreetMap Foundation (OSMF) for some frivolous reason, such as “my client used OSM maps on his CartoCrate Corp GPS Routing App on his iPhone and crashed into a Yak. OSM Maps are at fault.” This will cause a crisis in the Foundation, causing a mapping personality to raise a crowdfunding drive to raise legal funds, and in the process, the personality devolves the OSMF board and takes executive control. As a response to the legal mess and the insinuation pointed at them from some mapping extremists that they were involved, CartoCrate Corp will create a Public Domain fork of the OSM database and set about getting the US Govt using it. It’s quite successful.
- Oh, BTW Mapbox ended up as the company that did the paying mappers idea, although not as large scale.
Bill Dollins, Senior Vice President, Zekiah Technologies, Inc.
I think 2015 will be the year that other industries will catch on to geo and realize they don’t need GIS to do it. It won’t happen overnight, but I think it may be the start of a trend.
It could take any number of forms. Maybe it’s a company or organization not previously associated with geo that releases a narrowly-targeted tool or library that solves an industry-specific set of geo problems. Maybe it’s a large company acquiring a plucky geospatial startup to add some location awareness to its products.
The point is that the variety of open tools available will make it easier for organizations that understand their markets well to plug in “just enough geo” to enhance their domain-specific needs.
In short, 2015 will start to expose the fact that “Enterprise GIS” is really only relevant to GIS enterprises.
“The best way to predict the future is to invent it.” –Alan Curtis Kay
A year ago I predicted the beginnings of artisanal satellite mapping and ended with the phrase: “OpenDroneMap anyone?” After writing the prediction, I felt there was only one way forward. I registered the domain, proposed the project, and a year later (with the help and enthusiasm of many), we have OpenDroneMap. By January 2015, we will have a fully open source toolchain which fills 90% of the niche that Pix4D, Agisoft, and other closed-source projects fill in turning drone photos into geographic data, and it will have advantages that the closed source solutions know not.
Drones will continue to grow and be relevant to GeoHipsters in 2015 — more so in fact. As a result, retrograde coordinate systems, such as ones with promises of equal area, sub-inch accuracy, as well as vertical and horizontal datums galore will rise in prominence again amongst the GeoHipster clans. Geo will continue to converge to engineering scales. Great mock battles will be enacted between GeoHipsters and Survey-Hipsters.
But the great problems to solve in 2015 become ones not of how to process the new data, but how to store, retrieve, and share the data. In short, we will need artisanal pixel platforms like OpenAerialMap, with the need for similar projects specific to elevation models and digital surface models. These will be the great platform problems of GeoHipsters in 2015.
Companies exiting the tinkering phase will go all-in on making the killer apps, and building people-infrastructure like education and support. The GDAL/PostGIS/Mapnik stack that’s common to pretty much everything will see some major competition that fits better with the threads, clouds, and GPUs of new tech. Vector maps will finally be broadly available and the client side will eat more of the server side stack.
So my 2014 predictions were more wrong than right. Maybe all wrong.
OSM did get more national attention and has a lot of companies leaning on it now. I would still argue that the “program” as a whole is unprepared for the World Spotlight. The OSM Foundation rules by not ruling and it can still be a bit of a wild wild west shootout on the listserves… since I was just in one. For me HOT (Humanitarian OSM) is the most desirable component of OpenStreetMap. They are doing it right — and maybe that is what saves OpenStreetMap.
I still think more organizations are moving to hybrid setups. It’s the future. You can’t rely on one software ecosystem to provide all your needs and wants.
I thought Esri would take out MapBox… and I’m so glad they didn’t pull a GeoCommons.
The LAS format dust-up between Esri and rapidlasso finally gets some air time. I think this instance of Esri forking a LPGL format to only run on their software shows exactly how much they are embracing the open source world. They aren’t. If it happens once it will happen again.
Esri starts pushing the pay-for-play option for everyone. It may not be a bad thing at all. You pay for the portions of the ArcGIS ecosystem you want to use by buying “Esri Credits”. Hopefully that push finally kills off the out-of-date three-tiered approach to software sales. For some it will be good — for some it will be bad.
QGIS is my new love long-term affair for desktop software. I think in 2015 it takes off and becomes a tool everyone has on their desktop. Take a portion of your commercial maintenance (I’m guilty for not) and donate it to the program.
Esri makes a play for Spatial Networks. The guys behind Fulcrum are doing everything right. I have developed a severe bromance for that company. I hope I’m wrong on this one.
Finally — all the #geo people stop putting #geo in front of every term. Now I’m going out for a #geobeer to the #geopub and #georemember the one #geoprediction I didn’t #geomake.
DRONES. Which of course sounds hackneyed and outdated, but UAV capabilities keep leaping forward. In 2014 the hardware matured and reached a broad audience; for 2015 UAV software and data management seems poised to shorten the channel from lens to actionable data. I see it already in disaster response, but I think agriculture is the likely sector to blow the doors off of UAV capabilities.
When asked to look into the future, I generally start by looking into the past and reasoning by analogy. There have been some patterns of development in the last year that are worth at least commenting on.
As it happens, I saw this movie the last time around, when the Java community arrived in the early 2000s and rewrote all of geospatial. The dominant architectural paradigm of the time was the three-tier, built on open standards, and the software all shows it. Everything was XML-configured, everyone followed the Gang of Four, everything was multi-threaded, and everything Had-To-Be-In-Java.
Imagery — Cheap, Cheaper, Cheapest
For a long time, discussion of imagery has been pretty boring stuff: Can you convert formats? Can you ortho-rectify? Can you color-balance and feather a mosaic? But this year at FOSS4G there was quite an interesting collection of talks all pointing towards a blossoming new world of cheap, cheap imagery.
At FOSS4G 2014 in Portland I saw talks by Frank Warmerdam on processing imagery for Planetlabs, and by Aaron Racicot and Stephen Mather on mapping with cheap drone hardware. With cheaper and cheaper sensors — ranging from the shoebox-sized “doves” from PlanetLabs to the sub-thousand-dollar quadcopters — we can and will pile up larger and larger collections of raw imagery. This pile of imagery will in turn drive innovation in image processing software to convert it all into a rationalized view of the world.
The new world of cheap, cheap imagery is enabled by two parallel innovations: really really cheap sensors that produce generally inferior imagery but incredible volumes of it; and really really clever software that can leverage multiple images of the same object to infer extra information, like 3D models, best pixels and so on.
It was only a handful of years ago that Microsoft’s Photosynth technique of building a 3D model out of an unorganized collection of photographs was a computing marvel. That capability is now available on smartphones. At one time the best computer vision software was only available as proprietary licensed SDKs: now it’s all open source.
With cheap sensors and algorithms capable of dealing with the imagery they generate, we’re not far away from a new world of earth observation — both from orbit and from only a few hundred feet above.
I expect that the tools for processing raw drone imagery will only get better, and that once (if?) the FAA sets reasonable rules for use of drones in the civilian sector, the USA will be awash in drones mapping every corner of the country — for the local county, electric company, and real estate board.
More Spatial, Fewer Maps
Over the last couple years I’ve been using the words “spatial IT” a lot, as a description for a trend in the geospatial world: our previously specialized tools are no longer specialized, they are just add-on features to general purpose IT tools.
So databases have a spatial column type. Document indexing systems (like ElasticSearch) have a spatial search capability. “Geocoding” functionality is now built into almost every application that might happen to have an address field.
So IT professionals are now capable of delivering “GIS” results, without “GIS” software. The classic “notification report” of all houses within 100 yards of a re-zoning application is now just a tabular query-and-mailmerge operation. No map, no GIS.
Data visualizations, having run through a brief map-mania after the advent of Google Maps, are now coming back full-circle to a realization that sometimes the best map is no map at all.
Non-spatial actions — like running a web search — are setting up subtle spatial queries in the background: your Google search returns results that are “relevant” to you not just categorically, but also geographically. Just a list of results, no map, but more spatial than before.
And spatial actions, like asking for directions, are returning way-finding results. A list of directions, rather than a schematic. The bar for a useful result from a routing application has moved “up” — away from a schematic cartographic result, to a natural language description of a step-wise route.
I guess it’s no real surprise that, as the wider world moves away from maps and towards embedded spatial everywhere, in our own field there is a growing nostalgia for classic cartography. Who is winning the maps competitions at the spatial gatherings these days? The whooshing pulsing arrows of the data visualizers, or the clear classic cartographers, telling story with the spare language of maps?
Hopefully this is just a moment, and maps will be back, but for now, spatial is the thing.
Two major trends:
- More and more consumer services/apps will emerge that assume/require continual knowledge of the consumer’s location. This whole new range of services will require more and more consumers, but also developers, to contemplate location in ways they never have before. Privacy will be a bigger issue than ever. Most of the developers building these services will have no background in GIS or cartography or such things, nor any desire to learn. They will embrace whatever tools help them get the job done.
- OpenStreetMap’s data volume will continue to grow rapidly, not least due to the introduction of more and more domain-specific editors like Rob Hawkes’ building height estimator or Tom MacWright’s CoffeeDex. More and more data, all around the world, will flow in, in some cases without the consumer even needing to be aware OSM is the underlying datastore. Meanwhile on the usage side, OSM’s “good enough” quality will continue to improve every day. The inevitable march forward will continue, helped on by more and more governments (local, national) embracing open data albeit in a very chaotic, piecemeal manner. More and more best practices and robust toolchains will emerge.
More Geospatial Visualisations, Maybe Less Maps
One of the great things about having a wife who understands and accepts that you’re a map nerd is getting great Christmas gifts such as “London: The Information Capital” by James Cheshire and Oliver Uberti. The book is subtitled “100 maps and graphics that will change how you view the city”. Reading this book made me realise that there were as many maps as there were visualisations with spatial data, and I also realised that this book wasn’t an isolated instance. More and more geospatial information is being visualised, both online and elsewhere. This means lots of maps but not just maps. This is a trend that will continue into 2015 and beyond as people who aren’t used to maps still want to visualise mapping data.
More Tangible Maps
At home I’ve got gift wrap with maps on it, a notebook covered in maps and even a map on the case of my phone. Walking through our local bookstore at the weekend, I was struck by just how many maps there were in so many shapes, sizes and forms, and with not a single digital map to be seen. Take a brief search through Etsy and you can get a map on all you ever wanted and a lot more besides. Maybe the public has fallen back in love with tangible maps as digital maps become more and more part of our daily lives? Whatever the reason, maps are here to stay.
More Bad Maps
Both of the previous two predictions means we’re going to continue seeing a lot more maps. But that also widens the scope for a lot more bad maps. There’s even a Twitter hashtag for this. Take a search for #badmaps if you want your eyes to bleed and whatever cartographic skill you possess to shriek out in anguish. This is not going to get any better. Now, because anyone can make a map, this means anyone might make a map, regardless of whether they should or not. That’s not to say that cartography should remain the preserve of professional cartographers, but if you don’t have a modicum of appreciation of design, an eye for colours that complement each other, and at least a rudimentary understanding of geography, then making a map might be something you want to pause and think about.
Even More People Doing GIS, Without Knowing They’re Doing GIS
OSM Will Explode
It’s a sweeping generalisation and probably a controversial one too, but OpenStreetMap seems to be divided into 4 tribes. Firstly there’s the utopian tribe, truly believing that OSM is the only way forward for mapping data, that it will dominate across all forms of mapping data, and that if only everyone else would embrace the ODbL and its share-alike clause everything would be so much easier. Then there’s the community tribe, who use and contribute to OSM because they like the community aspect first and the mapping aspect second. Thirdly there’s the map tribe who just want to get on with mapping the world. Finally there’s the pragmatic tribe who want to see OSM flourish in the current business world and realise that something probably has to change in order for that to happen.
Each tribe wants something different from OSM and although there’s overlaps and blurring the lines, the OSM community is a divided one. I have to ask if this is sustainable in its current form.
All of which means that 2015 might be the year OSM explodes. Sadly this doesn’t mean uptake and contributions will explode, but my fearful prediction is that OSM itself will explode and fragment, with the possibility of OSM forking looming on the horizon.