[HOT] Questions regarding metrics from taskprojects, tasks and volunteers.
andrew.r.buck at gmail.com
Fri Jun 5 00:27:40 UTC 2015
-----BEGIN PGP SIGNED MESSAGE-----
See responses below...
> Can any of you help me with the following questions:
> 1. In pre-crisis taskprojects; Do you think it is important to have
> a person or organisation publishing announcements and "attracting"
> volunteers to a task?
The more people hear about the need to have mapping work done the more
people will respond to it and contribute. Even more important though
is that the request for help come directly from them. Most people
have never heard of OSM or HOT so it is difficult for us to convey to
people that the data really will be used to help people. However when
people hear directly from the Red Cross that the data is going to be
used, that sees to be the kind of thing that really gets volunteers to
More importantly, is a somewhat detailed explanation of how the data
will be used. Something like, "we are responding to flooding that
occurred along the ... river, we would like you to map the buildings
in the area so we can then use low resolution (but current) satellite
imagery to determine the exact number and location of flooded houses
in the area". This explanation of the need for the data not only
shows them you do have a concrete plan and as soon as this is done you
will do "some thing" and then use that for some purpose in your
overall mission. This is such a strong motivator for volunteers
(which are the driving force behind HOT) that one of the key criteria
for whether or not to begin an official activation is whether we have
gotten a "specific request" for data from a recognized aid
organization such as the Red Cross. Without such a request, the only
people who will end up working on the project is the small group of
people who map between more major disasters and this group is then
split up over the numerous small things or previous events that are in
the task manager at any given time.
So in answer to your question, the importance of the map data is what
determines this. Having someone actively working on writing blog
posts, talking to journalists, talking to the OSM/HOT community and so
forth about what the data will be used for will greatly increase the
number of volunteers who participate. Not only this, but it helps the
HOT project in an ongoing manner, because it increases people's
confidence in the work that we are doing.
> 2. Is there a way to get (a rough estimate) on time volunteers
> spend on a task?
This could be generated from the task manager. When someone locks a
tile they either unlock it, mark it complete, or after 2 hours it
automatically unlocks. For the fraction that are manually unlocked by
the user it would be easy to generate the time the tile was locked
for; not a perfect number for time spent, but ok for a rough estimate.
Also, since we know the users OSM username and the time they started
mapping, we can look into the history of the OSM database and see what
objects were added/modified by them around this time so we can see not
only how long they worked but what they did in that time. The OSM
history stats would be more difficult to generate, but it is possible.
> 3. Is there a (rough estimate) on the number of objects (nodes,
> ways, relations) an average volunteer produces during a task?
Answered in the previous question.
> 4. Do you think it would be possible, given the size of a
> geographic "area of interest" to estimate how many volunteers
> and/or mapping days would be required to succesfully complete a
This turns out to be very difficult to do. Not only is the number of
volunteers who contribute highly variable over time, there is also the
confounding effect of what other projects/activations are occurring at
the same time (and this is of course always changing). This means
that even looking at how long it took to map other projects in the
past doesn't give a complete picture because not only might effort be
being diverted to other projects than the one you are examining, but
also that the public's general awareness of the HOT project itself
waxes and wanes over time. When lots of people show up to help map
the recent Earthquake they not only do that but some of them see other
projects that interest them and jump in to work on them. So large,
news making disasters not only soak up a lot of effort, they also
bring in people and re-activate dormant mappers in a very
unpredictable fashion. This is not to say that it is impossible to
put time estimates on project completion times, but that the error
bars on these estimates will sometimes be large enough to render the
estimate meaningless. We can provide rough estimates, but that is
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
-----END PGP SIGNATURE-----
More information about the HOT