Re: [HOT] Mapswipe - whats happening with the data?

[Dale Kunce]

I love the above example. Mapswipe is a great tool for large scale mapping
of the type that HOT and Missing Maps does often.

Heidelberg University has done some great early work to visualize the
results and projects in MapSwipe. This is still a work in progress but its
a great example of the Missing Maps consortium members all pitching in and
doing a little bit of work for a greater good, just like OSM.

http://mapswipe.geog.uni-heidelberg.de/

On Tue, Mar 28, 2017 at 9:00 AM, Violaine Doutreleau <
v_doutrel...@cartong.org> wrote:

> Hi all,
>
> Just a quick message to share my point of view : mapswipe allows way more
> people to take part into Missing Maps than OSM, even though it's not direct
> mapping...Anyone who doesnt't want or doesn't know how to map in OSM can
> take part in this effort. That's its best advantage and where it save
> mappers time...
>
> Bests
>
> Violaine
>
> Le 26/03/2017 à 17:58, john whelan a écrit :
>
> >It's definitely useful to save actual mapper's time. For validation I'm
> not sure it would be a big help.
>
> I must be missing something.  Mapswipe uses four passes and then we get
> different sized tiles out of it.  We still need an old fashioned mapper to
> map the tile.  Traditionally if a mapper thought the tile size was too big
> they would split it to give different sized tiles.
>
> How does it save mapper time?  Does that take into account the four passes
> that mapswipe uses first?
>
> On the validation side really all you need is to inspect the tile and say
> its OK. If the tile is good enough you don't need to add anything.  From a
> practical point of view a conventional validator could do it in one pass
> provided we trusted them.  You'd still want to run JOSM for things like
> duplicate segments, crossing ways etc but that's faster than having to do a
> visual inspection as well.
>
> Cheerio John
>
> On 26 March 2017 at 10:46, Jo  wrote:
>
>> It's definitely useful to save actual mapper's time. For validation I'm
>> not sure it would be a big help.
>>
>> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :
>>
>> Would we be better ahead by using something like mapswipe to validate?
>>> If a tile is OK there is no need to add to the map from the tool.  If three
>>> mappers tick it then I would say its good to go.
>>>
>>> The current system we use four mappers to mapswipe, then a conventional
>>> mapper to map followed by a validator.  Six passes to get a validated
>>> tile.  The other way would give you one in four passes three of which would
>>> be on smartphones.
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 09:01, john whelan  wrote:
>>>
 But does it address the concern about how much effort is expended
 compared to the value added?

 Cheerio John

 On 26 March 2017 at 08:44, Ralf Stephan  wrote:

> Search is your friend. Searching for Mapswipe yields eg
> http://tasks.hotosm.org/project/2521
> where clearly the tiles are preselected. Also the description: "The
> data are prepared by MapSwipe"
>
> Regards,
>
> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
> wrote:
>
>> I seem to recall the same area is mapswiped by four different people
>> before a mapper maps it.
>>
>> Other than the size of the tile or which area to give priority to I
>> think the advantage is that people can do it anywhere on a smartphone.
>>
>> I would hesitate to say if it is worth doing or not.  In future there
>> might be a way to bring something into OSM via a review process.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 08:22, Hakuch  wrote:
>>
>>
>>
>> On 26.03.2017 14 <26.03.2017%2014>:13, john whelan wrote:
>> > My understanding is that mapswipe is only used to identify where to
>> map.
>> > Not to contribute to the map.
>> >
>> > Cheerio John
>>
>> of course there is no data transmitted to OSM, but how is the data (I
>> mean, marked tiles) used after I transmitted it? How much is it worth
>> to
>> do that mapswiping?
>>
>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>

>>>
>>> ___
>>> HOT mailing list
>>> HOT@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/hot
>>>
>>>
>
>
> ___
> HOT mailing 
> listHOT@openstreetmap.orghttps://lists.openstreetmap.org/listinfo/hot
>
>
> --
> *Violaine Doutreleau*
> Coordinatrice Missing Maps
> CartONG
> mobile : 06.95.02.42.44
> skype : doutreleau.violaine
>
>
> *P Help save paper - do you need to print this email?*
>
>
> --
> [image: Logo AVG] 
>
> Cet email a fait l'objet d'une analyse antivirus par le logiciel antivirus
> AVG.
> www.avg.com 

Re: [HOT] Mapswipe - whats happening with the data?

[Violaine Doutreleau]

Hi all,

Just a quick message to share my point of view : mapswipe allows way 
more people to take part into Missing Maps than OSM, even though it's 
not direct mapping...Anyone who doesnt't want or doesn't know how to map 
in OSM can take part in this effort. That's its best advantage and where 
it save mappers time...


Bests

Violaine


Le 26/03/2017 à 17:58, john whelan a écrit :
>It's definitely useful to save actual mapper's time. For validation 
I'm not sure it would be a big help.


I must be missing something.  Mapswipe uses four passes and then we 
get different sized tiles out of it.  We still need an old fashioned 
mapper to map the tile.  Traditionally if a mapper thought the tile 
size was too big they would split it to give different sized tiles.


How does it save mapper time?  Does that take into account the four 
passes that mapswipe uses first?


On the validation side really all you need is to inspect the tile and 
say its OK. If the tile is good enough you don't need to add 
anything.  From a practical point of view a conventional validator 
could do it in one pass provided we trusted them. You'd still want to 
run JOSM for things like duplicate segments, crossing ways etc but 
that's faster than having to do a visual inspection as well.


Cheerio John

On 26 March 2017 at 10:46, Jo > wrote:


It's definitely useful to save actual mapper's time. For
validation I'm not sure it would be a big help.

Op 26 mrt. 2017 3:12 p.m. schreef "john whelan"
mailto:jwhelan0...@gmail.com>>:

Would we be better ahead by using something like mapswipe to
validate?  If a tile is OK there is no need to add to the map
from the tool. If three mappers tick it then I would say its
good to go.

The current system we use four mappers to mapswipe, then a
conventional mapper to map followed by a validator.  Six
passes to get a validated tile.  The other way would give you
one in four passes three of which would be on smartphones.

Cheerio John

On 26 March 2017 at 09:01, john whelan mailto:jwhelan0...@gmail.com>> wrote:

But does it address the concern about how much effort is
expended compared to the value added?

Cheerio John

On 26 March 2017 at 08:44, Ralf Stephan mailto:gtrw...@gmail.com>> wrote:

Search is your friend. Searching for Mapswipe yields
eg http://tasks.hotosm.org/project/2521

where clearly the tiles are preselected. Also the
description: "The data are prepared by MapSwipe"

Regards,

On Sun, Mar 26, 2017 at 2:40 PM john whelan
mailto:jwhelan0...@gmail.com>>
wrote:

I seem to recall the same area is mapswiped by
four different people before a mapper maps it.

Other than the size of the tile or which area to
give priority to I think the advantage is that
people can do it anywhere on a smartphone.

I would hesitate to say if it is worth doing or
not.  In future there might be a way to bring
something into OSM via a review process.

Cheerio John

On 26 March 2017 at 08:22, Hakuch
mailto:hak...@posteo.de>> wrote:



On 26.03.2017 14 :13,
john whelan wrote:
> My understanding is that mapswipe is only
used to identify where to map.
> Not to contribute to the map.
>
> Cheerio John

of course there is no data transmitted to OSM,
but how is the data (I
mean, marked tiles) used after I transmitted
it? How much is it worth to
do that mapswiping?


___
HOT mailing list
HOT@openstreetmap.org 
https://lists.openstreetmap.org/listinfo/hot





___
HOT mailing list
HOT@openstreetmap.org 
https://lists.openstreetmap.org/listinfo/hot





___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot


--
Fwd: Re: Disponibilité de votre lieu pour un mapathon Missing Maps
*Violaine Doutreleau*
Coordinatrice Missing Maps
CartONG 

Re: [HOT] Mapswipe - whats happening with the data?

[joost schouppe]

Yes, Mapswipe is powerful indeed. Pete described the processing for Sierra
Leone. It didn't just result in "less work for the mapper", it also made
the mapping work itself possible for new mappers. We finished these tasks
(2664, 2680, 2710, 2711, 2712) before breakfast (well, almost) during our
Belgian National Mapathon this weekend. Almost all of them were new mappers.

In many tasks, previous mappers will have added some or all needed data,
but not mark the tile as done in the tasking manager. Or they look at a
huge patch of (African) satellite picture for the first time, and don't
find anything to map. This is all very confusing to a new mapper. In this
case, that was avoided, because we could focus on areas that we know were
empty AND have something to map.

There will be some write-ups about our big mapathon, and it might be worth
push-notifying something about how Mapswipe is used in practice to the
Mapswipe contributor, through the app. It is not enough to do cool stuff,
people need to really know they're making cool stuff possible :)

One thing that is up for improving, is that the preprocessing sometimes
results in miniscule tasks. It looks really strange and it is confusing for
users (an example: http://tasks.hotosm.org/project/2664#task/340). Just a
matter of using a slightly different processing. From what I see, I would
guess that the mapswipe work is used to create arbitrary polygons of
"please map me areas". Then they are split up according to some standard
grid, to make sure tasks aren't huge. That process needs some tweaking, for
example apply a rule like "if resulting task < 100 m², merge it to the
adjacent task".

-- 
Joost Schouppe
OpenStreetMap  |
Twitter  | LinkedIn
 | Meetup

___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

OK I accept that mapswipe is wonderful.

Thanks for the input.

Cheerio John

On 26 Mar 2017 12:52 pm, "Pete Masters"  wrote:

> Hey John, it's not just that its quicker. It is also that different people
> can engage with the task. It's hard to contribute if you have a few minutes
> through the tasking manager. But you can through MapSwipe. It's also near
> impossible to contribute via mobile or on the go. Again MapSwipe makes this
> possible.
>
> Even if it was as quick to do the scanning through JOSM, we are able to
> get a whole new group of people involved in the workflow.
>
> Cheers,
>
> Pete
>
> On 26 Mar 2017 17:27, "Jo"  wrote:
>
>> Hi John,
>>
>> Both possibilities exist. You can 'reserve' a bunch of tile to work on
>> while off line, or you can work online.
>>
>> It's easy to install it and play with it.
>>
>> Polyglot
>>
>> 2017-03-26 18:23 GMT+02:00 john whelan :
>>
>>> So what you're saying is it can scan about as fast as JOSM using
>>> .  Does it require an online connection or can it
>>> precharge so someone can do it on the bus and upload the results when it
>>> gets near WiFi?
>>>
>>> Thanks John
>>>
>>> On 26 March 2017 at 12:05, Pete Masters 
>>> wrote:
>>>
 Hi all, late to the party, but here are my two cents...

 MapSwipe, anecdotally for now, does save mappers' time. In an area
 where little OSM data exists, it can target mappers' efforts where people
 (or roads or rivers) exist. In the kind of rural landscapes we deal with a
 lot in MSF areas of interest, this cuts down the amount of time spent
 'scanning' for stuff to map significantly. And, MapSwipe users can get
 through a large area super fast.

 In the Sierra Leone projects we used it slightly differently. We knew
 that Kenema district had been partly mapped through several other
 initiatives (Ebola mapping  and the American Red Cross West Africa project
 for example). We wanted comprehensive data in OSM, so we tasked the area in
 MapSwipe and cross referenced the results with existing OSM data. Where
 MapSwipers said 'yes, there is a building' but none existed in OSM the area
 would be tasked for mapping. Where MapSwipers said 'yes, there is a
 building' and building data existed in OSM, we discounted the area from the
 tasking. Meaning that mappers were applying affords only where there were
 gaps in the data.

 I hope that is of interest. Couple of related points...

 Each area is classified by three different users in MapSwipe, not four.

 Paul Stewart wrote up the process for using Mapswipe to deduplicate
 mapping efforts through the tasking manager
 http://www.openstreetmap.org/user/pjstewart1984/diary/40710

 Disastermappers write here about how to access MapSwipe data and how
 they process it https://disastermappers.wordpr
 ess.com/2016/11/16/integrating-mapswipe-and-hot-tasking-mana
 ger-first-task-online/

 HOT Summit presentation on MapSwipe and pybossa
 https://youtu.be/pRZ_mWn0Lmc

 Cheers,

 Pete





 On 26 Mar 2017 15:47, "Jo"  wrote:

> It's definitely useful to save actual mapper's time. For validation
> I'm not sure it would be a big help.
>
> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan"  >:
>
>> Would we be better ahead by using something like mapswipe to
>> validate?  If a tile is OK there is no need to add to the map from the
>> tool.  If three mappers tick it then I would say its good to go.
>>
>> The current system we use four mappers to mapswipe, then a
>> conventional mapper to map followed by a validator.  Six passes to get a
>> validated tile.  The other way would give you one in four passes three of
>> which would be on smartphones.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 09:01, john whelan  wrote:
>>
>>> But does it address the concern about how much effort is expended
>>> compared to the value added?
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>>>
 Search is your friend. Searching for Mapswipe yields eg
 http://tasks.hotosm.org/project/2521
 where clearly the tiles are preselected. Also the description: "The
 data are prepared by MapSwipe"

 Regards,

 On Sun, Mar 26, 2017 at 2:40 PM john whelan 
 wrote:

> I seem to recall the same area is mapswiped by four different
> people before a mapper maps it.
>
> Other than the size of the tile or which area to give priority to
> I think the advantage is that people can do it anywhere on a 
> smartphone.
>
> I would hesitate to say if it is worth doing or not.  In future
> there might be a way to bring something into OSM via a review process.
>
> Cheerio John
>

Re: [HOT] Mapswipe - whats happening with the data?

[Pete Masters]

Hey John, it's not just that its quicker. It is also that different people
can engage with the task. It's hard to contribute if you have a few minutes
through the tasking manager. But you can through MapSwipe. It's also near
impossible to contribute via mobile or on the go. Again MapSwipe makes this
possible.

Even if it was as quick to do the scanning through JOSM, we are able to get
a whole new group of people involved in the workflow.

Cheers,

Pete

On 26 Mar 2017 17:27, "Jo"  wrote:

> Hi John,
>
> Both possibilities exist. You can 'reserve' a bunch of tile to work on
> while off line, or you can work online.
>
> It's easy to install it and play with it.
>
> Polyglot
>
> 2017-03-26 18:23 GMT+02:00 john whelan :
>
>> So what you're saying is it can scan about as fast as JOSM using
>> .  Does it require an online connection or can it
>> precharge so someone can do it on the bus and upload the results when it
>> gets near WiFi?
>>
>> Thanks John
>>
>> On 26 March 2017 at 12:05, Pete Masters 
>> wrote:
>>
>>> Hi all, late to the party, but here are my two cents...
>>>
>>> MapSwipe, anecdotally for now, does save mappers' time. In an area where
>>> little OSM data exists, it can target mappers' efforts where people (or
>>> roads or rivers) exist. In the kind of rural landscapes we deal with a lot
>>> in MSF areas of interest, this cuts down the amount of time spent
>>> 'scanning' for stuff to map significantly. And, MapSwipe users can get
>>> through a large area super fast.
>>>
>>> In the Sierra Leone projects we used it slightly differently. We knew
>>> that Kenema district had been partly mapped through several other
>>> initiatives (Ebola mapping  and the American Red Cross West Africa project
>>> for example). We wanted comprehensive data in OSM, so we tasked the area in
>>> MapSwipe and cross referenced the results with existing OSM data. Where
>>> MapSwipers said 'yes, there is a building' but none existed in OSM the area
>>> would be tasked for mapping. Where MapSwipers said 'yes, there is a
>>> building' and building data existed in OSM, we discounted the area from the
>>> tasking. Meaning that mappers were applying affords only where there were
>>> gaps in the data.
>>>
>>> I hope that is of interest. Couple of related points...
>>>
>>> Each area is classified by three different users in MapSwipe, not four.
>>>
>>> Paul Stewart wrote up the process for using Mapswipe to deduplicate
>>> mapping efforts through the tasking manager
>>> http://www.openstreetmap.org/user/pjstewart1984/diary/40710
>>>
>>> Disastermappers write here about how to access MapSwipe data and how
>>> they process it https://disastermappers.wordpr
>>> ess.com/2016/11/16/integrating-mapswipe-and-hot-tasking-mana
>>> ger-first-task-online/
>>>
>>> HOT Summit presentation on MapSwipe and pybossa
>>> https://youtu.be/pRZ_mWn0Lmc
>>>
>>> Cheers,
>>>
>>> Pete
>>>
>>>
>>>
>>>
>>>
>>> On 26 Mar 2017 15:47, "Jo"  wrote:
>>>
 It's definitely useful to save actual mapper's time. For validation I'm
 not sure it would be a big help.

 Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" >>> >:

> Would we be better ahead by using something like mapswipe to
> validate?  If a tile is OK there is no need to add to the map from the
> tool.  If three mappers tick it then I would say its good to go.
>
> The current system we use four mappers to mapswipe, then a
> conventional mapper to map followed by a validator.  Six passes to get a
> validated tile.  The other way would give you one in four passes three of
> which would be on smartphones.
>
> Cheerio John
>
> On 26 March 2017 at 09:01, john whelan  wrote:
>
>> But does it address the concern about how much effort is expended
>> compared to the value added?
>>
>> Cheerio John
>>
>> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>>
>>> Search is your friend. Searching for Mapswipe yields eg
>>> http://tasks.hotosm.org/project/2521
>>> where clearly the tiles are preselected. Also the description: "The
>>> data are prepared by MapSwipe"
>>>
>>> Regards,
>>>
>>> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
>>> wrote:
>>>
 I seem to recall the same area is mapswiped by four different
 people before a mapper maps it.

 Other than the size of the tile or which area to give priority to I
 think the advantage is that people can do it anywhere on a smartphone.

 I would hesitate to say if it is worth doing or not.  In future
 there might be a way to bring something into OSM via a review process.

 Cheerio John

 On 26 March 2017 at 08:22, Hakuch  wrote:



 On 26.03.2017 14:13, john whelan wrote:
 > My understanding is that mapswipe is only used to identify where
 to map.
 > Not to contribute to the map.
 >

Re: [HOT] Mapswipe - whats happening with the data?

[Jo]

Hi John,

Both possibilities exist. You can 'reserve' a bunch of tile to work on
while off line, or you can work online.

It's easy to install it and play with it.

Polyglot

2017-03-26 18:23 GMT+02:00 john whelan :

> So what you're saying is it can scan about as fast as JOSM using
> .  Does it require an online connection or can it
> precharge so someone can do it on the bus and upload the results when it
> gets near WiFi?
>
> Thanks John
>
> On 26 March 2017 at 12:05, Pete Masters 
> wrote:
>
>> Hi all, late to the party, but here are my two cents...
>>
>> MapSwipe, anecdotally for now, does save mappers' time. In an area where
>> little OSM data exists, it can target mappers' efforts where people (or
>> roads or rivers) exist. In the kind of rural landscapes we deal with a lot
>> in MSF areas of interest, this cuts down the amount of time spent
>> 'scanning' for stuff to map significantly. And, MapSwipe users can get
>> through a large area super fast.
>>
>> In the Sierra Leone projects we used it slightly differently. We knew
>> that Kenema district had been partly mapped through several other
>> initiatives (Ebola mapping  and the American Red Cross West Africa project
>> for example). We wanted comprehensive data in OSM, so we tasked the area in
>> MapSwipe and cross referenced the results with existing OSM data. Where
>> MapSwipers said 'yes, there is a building' but none existed in OSM the area
>> would be tasked for mapping. Where MapSwipers said 'yes, there is a
>> building' and building data existed in OSM, we discounted the area from the
>> tasking. Meaning that mappers were applying affords only where there were
>> gaps in the data.
>>
>> I hope that is of interest. Couple of related points...
>>
>> Each area is classified by three different users in MapSwipe, not four.
>>
>> Paul Stewart wrote up the process for using Mapswipe to deduplicate
>> mapping efforts through the tasking manager
>> http://www.openstreetmap.org/user/pjstewart1984/diary/40710
>>
>> Disastermappers write here about how to access MapSwipe data and how they
>> process it https://disastermappers.wordpress.com/2016/11/16/integrating
>> -mapswipe-and-hot-tasking-manager-first-task-online/
>>
>> HOT Summit presentation on MapSwipe and pybossa
>> https://youtu.be/pRZ_mWn0Lmc
>>
>> Cheers,
>>
>> Pete
>>
>>
>>
>>
>>
>> On 26 Mar 2017 15:47, "Jo"  wrote:
>>
>>> It's definitely useful to save actual mapper's time. For validation I'm
>>> not sure it would be a big help.
>>>
>>> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :
>>>
 Would we be better ahead by using something like mapswipe to validate?
 If a tile is OK there is no need to add to the map from the tool.  If three
 mappers tick it then I would say its good to go.

 The current system we use four mappers to mapswipe, then a conventional
 mapper to map followed by a validator.  Six passes to get a validated
 tile.  The other way would give you one in four passes three of which would
 be on smartphones.

 Cheerio John

 On 26 March 2017 at 09:01, john whelan  wrote:

> But does it address the concern about how much effort is expended
> compared to the value added?
>
> Cheerio John
>
> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>
>> Search is your friend. Searching for Mapswipe yields eg
>> http://tasks.hotosm.org/project/2521
>> where clearly the tiles are preselected. Also the description: "The
>> data are prepared by MapSwipe"
>>
>> Regards,
>>
>> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
>> wrote:
>>
>>> I seem to recall the same area is mapswiped by four different people
>>> before a mapper maps it.
>>>
>>> Other than the size of the tile or which area to give priority to I
>>> think the advantage is that people can do it anywhere on a smartphone.
>>>
>>> I would hesitate to say if it is worth doing or not.  In future
>>> there might be a way to bring something into OSM via a review process.
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 08:22, Hakuch  wrote:
>>>
>>>
>>>
>>> On 26.03.2017 14:13, john whelan wrote:
>>> > My understanding is that mapswipe is only used to identify where
>>> to map.
>>> > Not to contribute to the map.
>>> >
>>> > Cheerio John
>>>
>>> of course there is no data transmitted to OSM, but how is the data (I
>>> mean, marked tiles) used after I transmitted it? How much is it
>>> worth to
>>> do that mapswiping?
>>>
>>>
>>> ___
>>> HOT mailing list
>>> HOT@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/hot
>>>
>>
>

 ___
 HOT mailing list
 HOT@openstreetmap.org
 https://lists.openstreetmap.org/listinfo/hot


>>> __

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

So what you're saying is it can scan about as fast as JOSM using
.  Does it require an online connection or can it
precharge so someone can do it on the bus and upload the results when it
gets near WiFi?

Thanks John

On 26 March 2017 at 12:05, Pete Masters  wrote:

> Hi all, late to the party, but here are my two cents...
>
> MapSwipe, anecdotally for now, does save mappers' time. In an area where
> little OSM data exists, it can target mappers' efforts where people (or
> roads or rivers) exist. In the kind of rural landscapes we deal with a lot
> in MSF areas of interest, this cuts down the amount of time spent
> 'scanning' for stuff to map significantly. And, MapSwipe users can get
> through a large area super fast.
>
> In the Sierra Leone projects we used it slightly differently. We knew that
> Kenema district had been partly mapped through several other initiatives
> (Ebola mapping  and the American Red Cross West Africa project for
> example). We wanted comprehensive data in OSM, so we tasked the area in
> MapSwipe and cross referenced the results with existing OSM data. Where
> MapSwipers said 'yes, there is a building' but none existed in OSM the area
> would be tasked for mapping. Where MapSwipers said 'yes, there is a
> building' and building data existed in OSM, we discounted the area from the
> tasking. Meaning that mappers were applying affords only where there were
> gaps in the data.
>
> I hope that is of interest. Couple of related points...
>
> Each area is classified by three different users in MapSwipe, not four.
>
> Paul Stewart wrote up the process for using Mapswipe to deduplicate
> mapping efforts through the tasking manager http://www.openstreetmap.org/
> user/pjstewart1984/diary/40710
>
> Disastermappers write here about how to access MapSwipe data and how they
> process it https://disastermappers.wordpress.com/2016/11/16/
> integrating-mapswipe-and-hot-tasking-manager-first-task-online/
>
> HOT Summit presentation on MapSwipe and pybossa
> https://youtu.be/pRZ_mWn0Lmc
>
> Cheers,
>
> Pete
>
>
>
>
>
> On 26 Mar 2017 15:47, "Jo"  wrote:
>
>> It's definitely useful to save actual mapper's time. For validation I'm
>> not sure it would be a big help.
>>
>> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :
>>
>>> Would we be better ahead by using something like mapswipe to validate?
>>> If a tile is OK there is no need to add to the map from the tool.  If three
>>> mappers tick it then I would say its good to go.
>>>
>>> The current system we use four mappers to mapswipe, then a conventional
>>> mapper to map followed by a validator.  Six passes to get a validated
>>> tile.  The other way would give you one in four passes three of which would
>>> be on smartphones.
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 09:01, john whelan  wrote:
>>>
 But does it address the concern about how much effort is expended
 compared to the value added?

 Cheerio John

 On 26 March 2017 at 08:44, Ralf Stephan  wrote:

> Search is your friend. Searching for Mapswipe yields eg
> http://tasks.hotosm.org/project/2521
> where clearly the tiles are preselected. Also the description: "The
> data are prepared by MapSwipe"
>
> Regards,
>
> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
> wrote:
>
>> I seem to recall the same area is mapswiped by four different people
>> before a mapper maps it.
>>
>> Other than the size of the tile or which area to give priority to I
>> think the advantage is that people can do it anywhere on a smartphone.
>>
>> I would hesitate to say if it is worth doing or not.  In future there
>> might be a way to bring something into OSM via a review process.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 08:22, Hakuch  wrote:
>>
>>
>>
>> On 26.03.2017 14:13, john whelan wrote:
>> > My understanding is that mapswipe is only used to identify where to
>> map.
>> > Not to contribute to the map.
>> >
>> > Cheerio John
>>
>> of course there is no data transmitted to OSM, but how is the data (I
>> mean, marked tiles) used after I transmitted it? How much is it worth
>> to
>> do that mapswiping?
>>
>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>

>>>
>>> ___
>>> HOT mailing list
>>> HOT@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/hot
>>>
>>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Pete Masters]

Hi all, late to the party, but here are my two cents...

MapSwipe, anecdotally for now, does save mappers' time. In an area where
little OSM data exists, it can target mappers' efforts where people (or
roads or rivers) exist. In the kind of rural landscapes we deal with a lot
in MSF areas of interest, this cuts down the amount of time spent
'scanning' for stuff to map significantly. And, MapSwipe users can get
through a large area super fast.

In the Sierra Leone projects we used it slightly differently. We knew that
Kenema district had been partly mapped through several other initiatives
(Ebola mapping  and the American Red Cross West Africa project for
example). We wanted comprehensive data in OSM, so we tasked the area in
MapSwipe and cross referenced the results with existing OSM data. Where
MapSwipers said 'yes, there is a building' but none existed in OSM the area
would be tasked for mapping. Where MapSwipers said 'yes, there is a
building' and building data existed in OSM, we discounted the area from the
tasking. Meaning that mappers were applying affords only where there were
gaps in the data.

I hope that is of interest. Couple of related points...

Each area is classified by three different users in MapSwipe, not four.

Paul Stewart wrote up the process for using Mapswipe to deduplicate mapping
efforts through the tasking manager
http://www.openstreetmap.org/user/pjstewart1984/diary/40710

Disastermappers write here about how to access MapSwipe data and how they
process it
https://disastermappers.wordpress.com/2016/11/16/integrating-mapswipe-and-hot-tasking-manager-first-task-online/

HOT Summit presentation on MapSwipe and pybossa https://youtu.be/pRZ_mWn0Lmc

Cheers,

Pete





On 26 Mar 2017 15:47, "Jo"  wrote:

> It's definitely useful to save actual mapper's time. For validation I'm
> not sure it would be a big help.
>
> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :
>
>> Would we be better ahead by using something like mapswipe to validate?
>> If a tile is OK there is no need to add to the map from the tool.  If three
>> mappers tick it then I would say its good to go.
>>
>> The current system we use four mappers to mapswipe, then a conventional
>> mapper to map followed by a validator.  Six passes to get a validated
>> tile.  The other way would give you one in four passes three of which would
>> be on smartphones.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 09:01, john whelan  wrote:
>>
>>> But does it address the concern about how much effort is expended
>>> compared to the value added?
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>>>
 Search is your friend. Searching for Mapswipe yields eg
 http://tasks.hotosm.org/project/2521
 where clearly the tiles are preselected. Also the description: "The
 data are prepared by MapSwipe"

 Regards,

 On Sun, Mar 26, 2017 at 2:40 PM john whelan 
 wrote:

> I seem to recall the same area is mapswiped by four different people
> before a mapper maps it.
>
> Other than the size of the tile or which area to give priority to I
> think the advantage is that people can do it anywhere on a smartphone.
>
> I would hesitate to say if it is worth doing or not.  In future there
> might be a way to bring something into OSM via a review process.
>
> Cheerio John
>
> On 26 March 2017 at 08:22, Hakuch  wrote:
>
>
>
> On 26.03.2017 14:13, john whelan wrote:
> > My understanding is that mapswipe is only used to identify where to
> map.
> > Not to contribute to the map.
> >
> > Cheerio John
>
> of course there is no data transmitted to OSM, but how is the data (I
> mean, marked tiles) used after I transmitted it? How much is it worth
> to
> do that mapswiping?
>
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>

>>>
>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

>It's definitely useful to save actual mapper's time. For validation I'm
not sure it would be a big help.

I must be missing something.  Mapswipe uses four passes and then we get
different sized tiles out of it.  We still need an old fashioned mapper to
map the tile.  Traditionally if a mapper thought the tile size was too big
they would split it to give different sized tiles.

How does it save mapper time?  Does that take into account the four passes
that mapswipe uses first?

On the validation side really all you need is to inspect the tile and say
its OK. If the tile is good enough you don't need to add anything.  From a
practical point of view a conventional validator could do it in one pass
provided we trusted them.  You'd still want to run JOSM for things like
duplicate segments, crossing ways etc but that's faster than having to do a
visual inspection as well.

Cheerio John

On 26 March 2017 at 10:46, Jo  wrote:

> It's definitely useful to save actual mapper's time. For validation I'm
> not sure it would be a big help.
>
> Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :
>
> Would we be better ahead by using something like mapswipe to validate?  If
>> a tile is OK there is no need to add to the map from the tool.  If three
>> mappers tick it then I would say its good to go.
>>
>> The current system we use four mappers to mapswipe, then a conventional
>> mapper to map followed by a validator.  Six passes to get a validated
>> tile.  The other way would give you one in four passes three of which would
>> be on smartphones.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 09:01, john whelan  wrote:
>>
>>> But does it address the concern about how much effort is expended
>>> compared to the value added?
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>>>
 Search is your friend. Searching for Mapswipe yields eg
 http://tasks.hotosm.org/project/2521
 where clearly the tiles are preselected. Also the description: "The
 data are prepared by MapSwipe"

 Regards,

 On Sun, Mar 26, 2017 at 2:40 PM john whelan 
 wrote:

> I seem to recall the same area is mapswiped by four different people
> before a mapper maps it.
>
> Other than the size of the tile or which area to give priority to I
> think the advantage is that people can do it anywhere on a smartphone.
>
> I would hesitate to say if it is worth doing or not.  In future there
> might be a way to bring something into OSM via a review process.
>
> Cheerio John
>
> On 26 March 2017 at 08:22, Hakuch  wrote:
>
>
>
> On 26.03.2017 14:13, john whelan wrote:
> > My understanding is that mapswipe is only used to identify where to
> map.
> > Not to contribute to the map.
> >
> > Cheerio John
>
> of course there is no data transmitted to OSM, but how is the data (I
> mean, marked tiles) used after I transmitted it? How much is it worth
> to
> do that mapswiping?
>
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>

>>>
>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Jo]

It's definitely useful to save actual mapper's time. For validation I'm not
sure it would be a big help.

Op 26 mrt. 2017 3:12 p.m. schreef "john whelan" :

> Would we be better ahead by using something like mapswipe to validate?  If
> a tile is OK there is no need to add to the map from the tool.  If three
> mappers tick it then I would say its good to go.
>
> The current system we use four mappers to mapswipe, then a conventional
> mapper to map followed by a validator.  Six passes to get a validated
> tile.  The other way would give you one in four passes three of which would
> be on smartphones.
>
> Cheerio John
>
> On 26 March 2017 at 09:01, john whelan  wrote:
>
>> But does it address the concern about how much effort is expended
>> compared to the value added?
>>
>> Cheerio John
>>
>> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>>
>>> Search is your friend. Searching for Mapswipe yields eg
>>> http://tasks.hotosm.org/project/2521
>>> where clearly the tiles are preselected. Also the description: "The
>>> data are prepared by MapSwipe"
>>>
>>> Regards,
>>>
>>> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
>>> wrote:
>>>
 I seem to recall the same area is mapswiped by four different people
 before a mapper maps it.

 Other than the size of the tile or which area to give priority to I
 think the advantage is that people can do it anywhere on a smartphone.

 I would hesitate to say if it is worth doing or not.  In future there
 might be a way to bring something into OSM via a review process.

 Cheerio John

 On 26 March 2017 at 08:22, Hakuch  wrote:



 On 26.03.2017 14:13, john whelan wrote:
 > My understanding is that mapswipe is only used to identify where to
 map.
 > Not to contribute to the map.
 >
 > Cheerio John

 of course there is no data transmitted to OSM, but how is the data (I
 mean, marked tiles) used after I transmitted it? How much is it worth to
 do that mapswiping?


 ___
 HOT mailing list
 HOT@openstreetmap.org
 https://lists.openstreetmap.org/listinfo/hot

>>>
>>
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

Would we be better ahead by using something like mapswipe to validate?  If
a tile is OK there is no need to add to the map from the tool.  If three
mappers tick it then I would say its good to go.

The current system we use four mappers to mapswipe, then a conventional
mapper to map followed by a validator.  Six passes to get a validated
tile.  The other way would give you one in four passes three of which would
be on smartphones.

Cheerio John

On 26 March 2017 at 09:01, john whelan  wrote:

> But does it address the concern about how much effort is expended compared
> to the value added?
>
> Cheerio John
>
> On 26 March 2017 at 08:44, Ralf Stephan  wrote:
>
>> Search is your friend. Searching for Mapswipe yields eg
>> http://tasks.hotosm.org/project/2521
>> where clearly the tiles are preselected. Also the description: "The data
>> are prepared by MapSwipe"
>>
>> Regards,
>>
>> On Sun, Mar 26, 2017 at 2:40 PM john whelan 
>> wrote:
>>
>>> I seem to recall the same area is mapswiped by four different people
>>> before a mapper maps it.
>>>
>>> Other than the size of the tile or which area to give priority to I
>>> think the advantage is that people can do it anywhere on a smartphone.
>>>
>>> I would hesitate to say if it is worth doing or not.  In future there
>>> might be a way to bring something into OSM via a review process.
>>>
>>> Cheerio John
>>>
>>> On 26 March 2017 at 08:22, Hakuch  wrote:
>>>
>>>
>>>
>>> On 26.03.2017 14:13, john whelan wrote:
>>> > My understanding is that mapswipe is only used to identify where to
>>> map.
>>> > Not to contribute to the map.
>>> >
>>> > Cheerio John
>>>
>>> of course there is no data transmitted to OSM, but how is the data (I
>>> mean, marked tiles) used after I transmitted it? How much is it worth to
>>> do that mapswiping?
>>>
>>>
>>> ___
>>> HOT mailing list
>>> HOT@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/hot
>>>
>>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Hakuch]

On 26.03.2017 15:01, john whelan wrote:
> But does it address the concern about how much effort is expended compared
> to the value added?

if I put effort in something I of course want to know how, where and by
whom it is used.


0x1E6B7645.asc
Description: application/pgp-keys
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

But does it address the concern about how much effort is expended compared
to the value added?

Cheerio John

On 26 March 2017 at 08:44, Ralf Stephan  wrote:

> Search is your friend. Searching for Mapswipe yields eg
> http://tasks.hotosm.org/project/2521
> where clearly the tiles are preselected. Also the description: "The data
> are prepared by MapSwipe"
>
> Regards,
>
> On Sun, Mar 26, 2017 at 2:40 PM john whelan  wrote:
>
>> I seem to recall the same area is mapswiped by four different people
>> before a mapper maps it.
>>
>> Other than the size of the tile or which area to give priority to I think
>> the advantage is that people can do it anywhere on a smartphone.
>>
>> I would hesitate to say if it is worth doing or not.  In future there
>> might be a way to bring something into OSM via a review process.
>>
>> Cheerio John
>>
>> On 26 March 2017 at 08:22, Hakuch  wrote:
>>
>>
>>
>> On 26.03.2017 14:13, john whelan wrote:
>> > My understanding is that mapswipe is only used to identify where to map.
>> > Not to contribute to the map.
>> >
>> > Cheerio John
>>
>> of course there is no data transmitted to OSM, but how is the data (I
>> mean, marked tiles) used after I transmitted it? How much is it worth to
>> do that mapswiping?
>>
>>
>> ___
>> HOT mailing list
>> HOT@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/hot
>>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Ralf Stephan]

Search is your friend. Searching for Mapswipe yields eg
http://tasks.hotosm.org/project/2521
where clearly the tiles are preselected. Also the description: "The data
are prepared by MapSwipe"

Regards,

On Sun, Mar 26, 2017 at 2:40 PM john whelan  wrote:

> I seem to recall the same area is mapswiped by four different people
> before a mapper maps it.
>
> Other than the size of the tile or which area to give priority to I think
> the advantage is that people can do it anywhere on a smartphone.
>
> I would hesitate to say if it is worth doing or not.  In future there
> might be a way to bring something into OSM via a review process.
>
> Cheerio John
>
> On 26 March 2017 at 08:22, Hakuch  wrote:
>
>
>
> On 26.03.2017 14:13, john whelan wrote:
> > My understanding is that mapswipe is only used to identify where to map.
> > Not to contribute to the map.
> >
> > Cheerio John
>
> of course there is no data transmitted to OSM, but how is the data (I
> mean, marked tiles) used after I transmitted it? How much is it worth to
> do that mapswiping?
>
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

I seem to recall the same area is mapswiped by four different people before
a mapper maps it.

Other than the size of the tile or which area to give priority to I think
the advantage is that people can do it anywhere on a smartphone.

I would hesitate to say if it is worth doing or not.  In future there might
be a way to bring something into OSM via a review process.

Cheerio John

On 26 March 2017 at 08:22, Hakuch  wrote:

>
>
> On 26.03.2017 14:13, john whelan wrote:
> > My understanding is that mapswipe is only used to identify where to map.
> > Not to contribute to the map.
> >
> > Cheerio John
>
> of course there is no data transmitted to OSM, but how is the data (I
> mean, marked tiles) used after I transmitted it? How much is it worth to
> do that mapswiping?
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Hakuch]

On 26.03.2017 14:13, john whelan wrote:
> My understanding is that mapswipe is only used to identify where to map.
> Not to contribute to the map.
> 
> Cheerio John

of course there is no data transmitted to OSM, but how is the data (I
mean, marked tiles) used after I transmitted it? How much is it worth to
do that mapswiping?


0x1E6B7645.asc
Description: application/pgp-keys
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[Hakuch]

On 26.03.2017 14:15, Jo wrote:
> Tasks.hotosm.org Sierra  Leone​ tasks were preprocessed using Mapswipe
> results

and in what way preprocessed?


0x1E6B7645.asc
Description: application/pgp-keys
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapswipe - whats happening with the data?

[john whelan]

My understanding is that mapswipe is only used to identify where to map.
Not to contribute to the map.

Cheerio John

On 26 Mar 2017 7:38 am, "Hakuch"  wrote:

> I contributed some data via mapswipe, but I could not find out where and
> how exactly the marked tiles are used? Are there tasks where the marked
> tiles are beeing mapped? Can I see somewhere, what and how much data I
> contributed via mapswipe?
>
> greets
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

[HOT] Mapswipe - whats happening with the data?

[Hakuch]

I contributed some data via mapswipe, but I could not find out where and
how exactly the marked tiles are used? Are there tasks where the marked
tiles are beeing mapped? Can I see somewhere, what and how much data I
contributed via mapswipe?

greets


0x1E6B7645.asc
Description: application/pgp-keys
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot