On Jun 16, 2010, at 2:54 PM, ext Sean Gorman wrote:
> Very few of the people deployed to deal with a disaster - earthquake, oil
> spill, hurricane are developers.
Very few of the people deployed to deal with a disaster have
any way of dealing with GIS data, in my estimation. Are you
saying you feel otherwise?
> The presence of a well documented API is not particularly helpful to most of
> the people responding.
No; it's helpful to making data available so that people
who know how to work with GIS data can turn it into something
that's actually useful.
> This is why when the World Bank wanted to provide open data in Haiti they
> sent terabyte thumb-drives with the raw data available, so it could be opened
> in multiple applications and repurposed, or if they did not have applications
> there was an OpenLayers viewer for the data. You could get the data on-line
> or off-line. If I'm out setting up a boom in Gulf I'm probably not going to
> have Internet access.
And who built that viewer?
> I think what we should be debating is what is the right open data strategy
> for disasters, not who knows ArcGIS server the best. I fully concede Chris
> will win this contest. If someone builds an AGS map that allows me to
> download the data from it so I can use it off-line and in other applications
> - giddy up that is awesome. My concern is the default I see in the field is
> an AGS "map service" that just gives me a picture. I don't find it terribly
> surprising that AGS has a query function for their server. Bit embarrassed I
> missed the link for it, but does not change the concern for the disaster
> responder trying to get data to do their job. A disaster responder who will
> likely be without Internet, not know GIS, not know Web development, and could
> greatly benefit from something better than a paper map or a PNG/PDF.
What exactly do you want here?
There are multiple problems that need to be looked at,
and eventually solved, for this and future disasters.
1. Data availability. All data gathered -- especially by
government organizations, but also by corporate entities,
if government is participating in recovery -- should be
made available to the public as quickly as is practical.
This type of task is the kind of thing that organizations --
governmental and otherwise -- are bad at. Publishing data,
and making people aware of it, are *hard* problems. (Ask
OSGeo, which has been bickering about Metadata for years.
Ask OAM, which died on the vine due to lack of love. Ask
anyone who has ever tried to create some kind of GIS data
catalog. I'm sure that some of the people in this discussion
understand this.)
Gathering data and providing it to the public should be a
high priority in all disaster scenarios.
2. Data usability. Once data has been gathered and provided, 90%
of the target users can't do what they need to do with it.
Take the haiti example: over 7 terabytes of imagery were made
available to the public. There were very few people who were in
a position to consume that data successfully.
Sean argues that providing the raw data is useful to many users
here; given the number of hits that the WMS got, vs. the raw
source got, I personally doubt this pretty significantly. The
number of users accessing via a friendly web interface is going
to be orders of magnitude more useful than someone who is handed
a 100GB TIFF and told "Go to it."
The 'raw data' that got shipped to haiti was, in part, not 'raw';
there was (in one case) an offline-friendly cache of tiles instead.
(Taking big images and turning them into tiles is something you can't
do easily on a thumb drive.) There probably was some raw data, but
feedback I got indicated that raw data was not the most useful.
I would argue the most useful thing to do is to help people use their
data. Putting it in an ArcGIS server is actually a step forward in
this direction because it produces maps (that people can use or
cache) and provides *some* UI for doing queries. It's not a perfect
UI, but it's a lot better UI than you get from a shapefile sitting
on your disk.
3. Data Evaluation + Derivation. In general, this is the part of the
puzzle that the 'general' community can help with, assuming #1 and #2
are solved. In the case of the Haiti earthquake, this role was played
by OpenStreetMap. This is the stage where non-experts may be able to
use published *and usable* data to create new sets of data.
OpenStreetMap users fell in love with WMS services during the Haiti
crisis. These WMS services were not providing raw data to users, but
were providing a processed data which could be used to help generate useful
secondary datasets, especially in the form of GPS devices and other
mapping interfaces used by responders.
The WMS is "nothing more than a PDF" -- according to Sean -- but was the
single most useful service to responders looking to help build a collection
of data to help respond to the Haiti disaster.
Each of these steps is done by a different group of people. The first has
to be done by data collectors, government agencies, etc. -- people who are
on the ground at the disaster site. (They can be amatuers, but must be
nearby to be effective.) Publishing the data from this is currently difficult;
improving this is something that the open data community really needs to
improve if there is a serious goal of having people use these solutions.
http://haiticrisismap.org/layers.html
Is an example of #1 for Haiti.
#2 is something that experts need to step up to do. Whether they are GIS
experts working to build summaries and reports, technology/web experts
making easy to use UIs available (http://haiticrisismap.org/), or
experts at some other field -- thisi s the step where the raw data
is transformed into something that the people who need i most can use.
#3 is something that people are talking about here: getting crowdsourced
data generated by accomplishing #2. Once it's in OSM, or other similar
tools, this type of thing can grow quickly with an active community.
(I will say that the ability to save lives, as well as the quickly available
high quality data, made the Haiti situation unique in this regard; the
motivation to help protect fishes, or even beaches, is lower than the
desire to help save *real people trapped in rubble.*)
Sean, I feel like you are saying that if #1 happens, #2 and #3 happen
magically. I've observed too many counter examples to believe that #2
gets done in a way that #3 will ever happen; the reason I stepped up
during the Haiti disaster was because I felt like there were a million
chickens running around with their heads cut off. Until the tools to
do what I personally did for Haiti are trivial for non-experts to use
and accomplish the same goals, this will simply not be the case.
#1 is hugely important. #2 is as well. For this disaster,
#1 is not being done well, nor is #2. But arguing that it's not being
done at all -- or that the fact that data is published is 'not good
enough' for some reason -- is just not fair to the people who are
putting at least some effort into getting some data published.
There is always room for improvement. That doesn't mean that discounting
efforts to try to do better is appropriate; we just need to show what
can be better, probably by example. The data behind these things --
such as it exists! -- appears to be at least available in part. Improving
the distribution of the data -- but more importantly, *doing #2 better*,
are the ideal way to show the world the way things really should be done.
All of this is in my opinion, of course. I'm usually wrong, but I
think that the end result of the efforts in mapping Haiti may show
that I'm not totally crazy.
Regards,
--
Christopher Schmidt
Nokia
_______________________________________________
Geowanking mailing list
[email protected]
http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
