Public Lands Stewardship Ethic and Linked Open Data
Backstory
Over the last few years, my hobbies and interests keep pointing me toward areas where we need to link a whole lot of related but functionally “unlinkable” data together. There are thousands of cases of little pieces of information about the same things scattered across databases and information stores of various kinds. In data and information sciences, these are often considered de-duplication challenges, but they are really more often proximity analysis and dot connecting problems. It’s not so much that we have actual duplicates of records about the same things but that we have many sources of properties or attributes about things that could be combined to create a fuller picture.
My latest challenge underlies a venture I’m pondering on some form of public lands stewardship database. I joined up recently with the Colorado Canyons Association’s Adopt-a-Trail program and started looking into where I could make contributions across the three National Conservation Areas in Western Colorado. I celebrated Juneteenth by taking my first couple of forays out to the Gunnison Gorge NCA and recording observations along the trails using Gaia GPS (my favored mobile mapping solution at the moment) complete with pictures and waypoint notes. Both ATV time and hiking time, picking up garbage/recyclables and pondering on being intentional in actively stewarding our public lands, got me thinking about what all our volunteer time doing this work might entail. Being a data guy, I also started thinking about how various adopt-a-trail volunteers might record and share data from these experiences.
I thought that two things would probably be true:
- There must be many more adopt-a-trail types of efforts all over the world.
- Someone has probably come up with some pretty cool databases and apps and whatever to capture details about activities like this.
On the adopt-a-trail idea, there are certainly quite a number of those in play, including one here in Colorado I’ve encountered before. Some of these are operated in the US by our land management agencies or their associated “friends-of” associations, and many of them focus on trail construction and more major maintenance efforts (sometimes in cases where a public agency is underfunded to pay for this work).
On the data and apps front, there are a couple of specifics for hiking and other trail activities that I already use:
- AllTrails is super basic with a rating/comment structure plus options for activity selection and trail conditions. I’ve actually found pointers to trails all over the world with AllTrails, so it’s scope is pretty large.
- COTrex is specific to Colorado and super nicely done as an app by the Colorado DNR and Parks and Wildlife. They have a more sophisticated “trip report” feature with the ability to record activity type, link an actual recorded route and selections on trail conditions, crowding, bugs, water, and bathrooms along with photos.
And I’m sure there are quite a few others as well, some of which might be worth exploring. Being a data guy who has also dedicated much of my career to open data and open science practices, I immediately start looking not just at what a particular app or platform has done but the underlying data situation and especially the openness or lack thereof for the data. AllTrails along with more GPS-centric platforms like Gaia GPS and onX are commercial ventures that are going to have a vested interest in protecting their unique underlying data assets. COTrex is a state government funded deal, and I did find a record in Colorado’s open data catalog that provides a snapshot of the underlying trails (and other) GIS data. This will probably be the same data that are incorporated into the USGS National Digital Trails Dataset (more on that later). Another outdoor activity mapping app, TrailForks, has an API that can be accessed with a key (I’ve requested one). Their trip report deal is a little more sophisticated than AllTrails but the focus is on uploaded pics and videos with a bent toward mountain biking at their roots. I don’t yet fully understand the use restrictions on their data.
None of these platforms or approaches gets at what I started dreaming up on the trail the other day. I was thinking about how an organization like the CCA or managing agencies like BLM could leverage data collected by their volunteers in various ways. As a simple example, there’s actually no way for me as a new volunteer with CCA to check out where other volunteers have already been working in our large coverage area (over 600 square miles between the three NCAs), when they were last there, and trends in overall trail messiness. The training materials for this job were well done, a joint effort between the CCA and BLM, but everything points to making phone calls to report anything in need of law enforcement or other professional attention on public lands. I can also imagine tying in with other platforms like Seek from iNaturalist for volunteers with an interest to record and share important biological observations like the couple areas I recorded invasive knapweed this week on the Chukar Trail. I also think about what a powerful open database this could be on a global scale for all kinds of research and for community momentum building.
The basic idea of the trip report has obviously been implemented in various ways, and I’m sure there are many other cases around the world where specific apps or data platforms have done something like this within their context. But I think we need something bigger and different than what I’ve seen so far. I’m imagining more of an open data framework based on linked data principles and technologies. We need to leverage many different platforms and tools for recording and sharing information, linking into and through them to develop intelligence about our trails and other public lands features. We need to think much more holistically about what it means to experience our public lands and join together in active care and stewardship.
The COTrex Trip Report is the most robust of those I’ve come across so far. In a slightly different context, FreeRoam for campgrounds and related sites is a similar idea with a little bit of a bent toward digital nomads with their social reporting feature. All of these are built within a particular context, collecting information that is specific to the app designers, what they want to use, and what they think users will spend time submitting. But what if we could link anything and have a much more extensible data schema with an ability to expand into limitless dimensions on what people might be willing and interested in sharing and what other people might be interested in learning about or analyzing.
Public Lands Stewardship Motivation
A Colorado Public Radio story with a guy named Nelson Holland really caught my attention a while back. Nelson talked about the dynamic of being a black hiker in what is still very much a white dominated activity and US State. He spoke in the story about the way he sometimes gets obviously weird (if not hostile) looks from especially older white hikers in the more rural parts of Colorado not used to seeing a black guy out on “their” trails. This struck me partly because I live in one of those parts of the State. I was thinking about that story when I was hiking down the Chukar Trail yesterday, the second officially Federally recognized Juneteenth. I was pondering the social aspects of public lands stewardship and thinking about what responsibilities and opportunities as an ally I might develop in promoting more inclusivity and equity. I was also thinking about the Nuchu (Ute) ancestors who may well have used the same route up from the Gunnison River before the lands and the wealth that came with them were stolen and swindled away. What can I do to promote and bring about more equity and inclusion in our public lands, and what reparations can I help lead toward for the damages wrought through colonization of what are now part of the public trust?
All of this has me thinking about what intentional and impactful use and stewardship of our public lands and their many natural and cultural resources might entail and how a linked data and information framework might support and be a part of making that more of a reality. Picking up garbage, hauling out recyclables, scooping up dog poop, and reporting on any issues in need of trail maintenance are all the basic jobs for adopting trails through our local association. Telling others about what we find and don’t find on the trails seems like it should be a part of that work. Perhaps spending a little time learning about a trail and its history and synthesizing that information in some way could also be useful. Acknowledging for myself and witnessing for others where opportunities for greater equity might be sought might also be important. I’m not sure how to bring this about, but I’ll be exploring it further.
Linked Data Challenge
While there are actually a lot of features on our public lands that we could essentially use as rallying points for a linked data framework (mountains, rivers, lakes, etc.), I’ll focus on trails as the problem du jour. The trails data situation basically sucks. As I mentioned, USGS has taken a crack at building a somewhat comprehensive approach to an ongoing data integration exercise with their open database. It’s not horrible, but it’s not that great either. It would be far better if they shifted their focus from building a new database to exposing what goes on in effectively linking or amalgamating multiple disparate sources, sharing the full depth of what is brought together so that other data builders can make their own choices about how to handle the issues and link more effectively back to original sources for different processing.
AllTrails uses OpenStreetMap as at least part of their source material (see this article on their derivation methodology). The bigger commercial players are opaque on sources, but there’s lots of user chatter indicating better or worse underlying data. I have a personal interest in contributing to and making OSM work better along with related efforts in the “wikiverse,” particularly Wikidata. Being open public data and providing APIs, I’ve started poking around at the intersections between the USGS National Digital Trails Dataset, OSM’s “highway=path” data, and Wikidata instances of trail. My thinking is that a combination of OSM and Wikidata should provide a fundamental platform upon which to build some type of globally scaled linked data architecture. Initial exploration, however, points to the need for a fair bit of robot and human curation work.
While geospatial data describing trails, including individual segments between sometimes other meaningful features, is reasonably robust based on the various ways it has been assembled and is maintained, the linkage to and information about the more abstract concept of a named trail is really quite lacking. AllTrails and various other groups have tried to address this issue based on their own requirements and available capital to build usable data, but those are not buildable as open architecture. Wikidata can be a powerful linked open data platform to rally around, including the practice of linking external identifiers together on the same entity. However, there is a current dearth of good records on trails pulled into Wikidata through a handful of bot actions. There are only 23 named and identified trails tied to Colorado, though there are some others that are just not appropriately attributed.
On the OSM side, there are 37,392 “paths” in Colorado, of which only 4,676 have unique names (much more work to be done here). Only three of those OSM ways have a wikidata tag pointing to a QID. Of those, the Southwest Ridge from Fourmile Creek “trail” on Mount Sherman in OSM links to the Wikidata entity for Mount Sherman. I would say that’s an inappropriate association recorded in OSM data, but the rules are a little ambiguous.
In my ideal world here, we’d have a Wikidata QID for every named trail to use as a digital lodestone. The Wikidata entity for the Canyon Rim Trail on the Colorado National Monument is perhaps a great exemplar; a very robust record that contains a wealth of linked information. The Wikiverse infrastructure would help to disambiguate between trails of the same name, and the Wikidata entities would provide a rallying point for any other identifiers for the same logical feature to help us link between disparate sources. But that world doesn’t exist yet, at all. So, the question is, can we start building that foundation? Can we build some bots to start assembling the stuff that can be verified, using the Wikidata conventions toward referencing everything to build a trustworthy platform? I can start to envision a preponderance of evidence approach to data integration that could develop enough clues about enough trails, at least within the US, to get the ball rolling in the right direction. We could then build some useful apps on that foundation and hopefully spark some interested human curation to help connect dots on the less certain stuff.
Public Trust Philosophy
There’s the geeky part of what I’m working toward that will be fun and challenging as yet another data integration problem to be solved with some creative software and technology. But I’m much more interested in promoting and doing what I can to support an ethic of stewarding and maintaining the lands and other resources in the public trust. I’m interested in providing an open data foundation that can serve as a freely available asset for groups like the CCA or the guide service startup that Nelson Holland has going. I’m interested in providing a focus for anyone who wants to engage with our public lands in whatever responsible and sustainable ways they want, being mindful of who has been there before, who will come after, and always working to leave the resources in the public trust in better shape than when they started.
