On Nov 2, 2018, at 9:31 AM, John Whelan <[email protected]> wrote:
> My feeling is OpenStreetMap has two sides. The first is local adding local
> knowledge to the map. The other I'll call armchair mapping. When Stats
> Canada did the pilot it tapped the local Ottawa mappers who meet physically.
Speaking from nearly a decade of experience, OSM has many, many sides, though
the two that John Whelan identifies are two many find "readily apparent." I
quick-read the study, which was actually quite informative in that it broke up
similar crowdsourcing efforts (OSM is only lightly mentioned) into demographic
categories, with some surprising results. One is that many volunteers are
older, sometimes disabled (stroke victims noting benefits of "repetitious tasks
which help my brain to heal" was cited) and have a particular need for the
sorts of social feedback which projects like this uniquely offer. (At the same
time, there is often a sharp dichotomy between these sorts of crowdsourced
projects and social media, with many in the study who prefer the former
appearing loathe to use the latter).
Another very important take-away is how participants in projects like these
truly improve their skill-sets (seriously improving quality of submitted data)
over time: like many things, the longer one participates, the better become
their skills. This emphasizes the importance of "growing experts," something
seldom mentioned in OSM.
> I would agree that amongst mappers with the most edits there is a high number
> of retired people and those with disabilities involved and these may not be
> visible. Tapping them for groups coming together to map can be a problem.
It might appear that way (that they are invisible), yet there is no denying
that "they find you." In short, "build a project that attracts older, likely
high-skill (or can grow there) participants, and they will come."
> In my view typically the most productive mappers are those with a special
> interest. Adding WiFi access or churches for example or even a change of
> street name.
While it is difficult to say why mappers become productive, it may be even
harder to do the apparently more simple task of defining "productive." I know
one mapper who flits about the entire planet in OSM, seeking to "up his stats
on a leaderboard" as he measures the number of edits he makes in the tens or
hundreds of thousands. Needless to say, the quality of his edits, and how
productive he is, is a matter of contention. There is such a thing as "high
quality" and in OSM this can and should be defined and refined especially for
major projects.
Certainly "quality of data entered" is one metric, yet even that can be hard to
define or measure, unless strict criteria are established at the beginning of a
project as a goal to strive. Once again, and especially in highly ambitious
projects (like BC2020) this underscores the need for some up-front planning,
up-front project management, up-front expectations of data quality and up-front
documentation of all of these things so that these expectations are met and
measured along the way. (Project Management 101, really).
> We also have a number of teachers who would like to use OSM and in particular
> the building project to involve their students. We get a fair amount of data
> added but the quality can be questionable. HOT and others I think have found
> that using a restricted set of tasks and tags works best.
I personally have experienced helping professors at the university level
(computer science, environmental studies...) use OSM, as students at the
undergraduate level readily take to OSM. Younger students (high school, middle
school) enjoy some success with it, more often at smaller, less ambitious
tasks, a recently popular one in the USA being "micro-mapping our local school
campus." (Drinking fountains/water stations, extremely detailed sporting
facilities, footways and associated potential routing, landscaping,
restricted/off-limits areas, parking areas for autos, motorcycles, bicycles,
etc.) What often works is breaking students into functional groups (sports
facilities, transportation, amenities...) and having a teacher/administrator
check the results of each group. The tags can start out restricted and stay
that way, or they can start out restricted and allow the students to develop
further "depth" by researching OSM's wiki pages, or even (yes, this is
advanced) structure their own scheme. For example, a high school has four
different libraries in several different buildings, or extensive sports
facilities, how might we best tag these?
And whether young, old or in-between, Martijn van Exel (an OSM superstar) has
proven with his (well, largely his) MapRoulette project that "gamification" can
really super-charge particular kinds of data entry/improvement sub-projects
like few other strategies can. The Study confirms this, saying "Platform
features such as gamification, quizzes and podcasts are frequently cited as key
enablers for many crowdsourcing campaigns." (Reed, Raddick, Lardner & Carney,
2013).
Another important aspect is intra-project communication, which in OSM's case
are things like wiki, talk- mailing lists such as this one, our forum to ask
questions and get answers, the built-in volunteer-to-volunteer "missive" system
allowing short messages to be exchanged and so on. A dangerous trend in OSM
has been towards "walled gardens" such as social media and tools like Slack,
which are proprietary and I (and others) have characterized as "secret-sauce
walkie-talkies." OSM doesn't benefit by those nor this trend and should avoid
anything but open platforms for intra-project communication.
> My personal feeling is giving feedback is useful. So the challenge for the
> building project is how to engage people. What are the most useful tags to
> add?
The Study stresses the importance of feedback. I agree this is an important
component of any crowdsourced project and In OSM one of the most successful
aspects of this is the near-immediacy with which the data recently entered get
rendered (assuming the data included tags which DO get rendered, not all do).
Identifying "useful" tags is part "what are the most important data this
project attempts to GET entered?" and part "how often and frequently-updated
are these data displayed after they ARE entered?" And a Tasking Manager should
be part of that for "more major" projects.
> I'd suggest some sort of web site giving the number of buildings mapped and
> the tags that have been added by city. Graphs with time as one axis would be
> nice.
While the study doesn't exactly say this about OSM, I'll offer that it is my
FIRM experience that using visual media like a map with blocks on it that show
color-coded progress (as Tasking Manager does) not only "stays within" the
geographic theme of mapping ("using a map to show how much we've mapped"), it
also is MUCH more visually appealing than graphs, tables or text-oriented data,
especially as these need to be scrolled or otherwise interacted with via
mouse/keyboard/other input device. One map, already displayed, with nothing
else needed except to look at it and "visually parse" what it is conveying,
really works well, perhaps even "works best." I have no hard data with which
to back that up, but I do have much personal experience that this is true.
> Certainly certain activities are more complex than others. Importing
> buildings is not a task I'd suggest for teenage mapper with twenty minutes
> experience. Breaking out the tasks is a task in itself and for 4 million
> buildings I think it could benefit from a project plan.
It will not simply "benefit" from a project plan, something as ambitious as a
national building project of the sort BC2020 aspires to be absolutely,
positively requires a project plan. I'm pretty sure the experiences we've had
here speak volumes about that.
> I think we've seen with the 2020 project that just saying it would be nice to
> have by is not really enough to sustain it but who would do it I'm not sure.
The initial approach of having federal government's imprimatur doesn't seem to
have worked out as it might have, but that doesn't mean it can't or shouldn't
play a role, I believe it can and likely should. It might make sense for
something like a STATSCAN or provincial-level agencies "acting as conductor,"
knowing best their data, crucially understanding how OSM has its own
methodologies for receiving data, and finding the "best blend" of making that
happen, publishing/promulgating a project plan for that to be crowdsourced in
an OSM-sensitive way, then largely stepping aside and letting the crowd do our
thing, likely offering (minor) "mid-course correction" if/as needed. This can
and does work, though it frequently benefits from identifying highly dedicated
and high-skills leadership within the crowd and directly supporting these
(usually quite few) people, enabling them with the ability to slightly modify
rules, offer seminars/educational curricula or even direct towards them
carefully-identified financial support to facilitate completion of the task.
There are many, many flavors of how this can happen, and it can often be most
useful to allow it to become "home grown" along the way, but with some firm
planning up-front to make sure it doesn't go awry (contingency plans,
worst-case-scenario anticipation...are crucial).
That's plenty for now!
SteveA
California
_______________________________________________
Talk-ca mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/talk-ca
