[HOT] TomNod integration for the next time around

Kevin Bullock kbullock at digitalglobe.com
Wed Apr 29 21:24:39 UTC 2015

Pierre, Bill, I can confirm the imagery source is the same. Meaning, the imagery we licensed openly was used both by HOT (via Mapbox) and on the Tomnod platform. I am actually putting a blog post together on this topic, and this thread has been *very insightful*. Barrier of entry is indeed one of the huge differentiators! As an example, my kids can participate on Tomnod, but need another decade or so before they can contribute to OSM ☺. As such, we are seeing (I believe) a higher volume of contributors via Tomnod: over 25,000 unique contributors so far.

So far, here is what the Tomnod team has validated, AND, here is a link to the GeoJSON:

488 Damaged Buildings Verified, 36 Damaged Roads Verified
Download the Validated geojson:  https://s3.amazonaws.com/tomnod-james/nepal_earthquake_2015_15448_validated.geojson.json.zip

Sincerely, Kevin

From: William Morris [mailto:wboykinm at geosprocket.com]
Sent: Wednesday, April 29, 2015 2:26 PM
To: Pierre Béland; hot at openstreetmap.org<mailto:hot at openstreetmap.org>
Subject: Re: [HOT] TomNod integration for the next time around

Hi Pierre,

I think we're already using the same imagery; it's entirely from DigitalGlobe, and the Mapbox crew processed that and made it available to us within hours of it being released (Camilla or Charlie - correct me if I'm wrong). The issue is that participants (~5,000 people this time) in DigitalGlobe's Tomnod project are deriving basic damage assessment data from the same imagery, and that information isn't making it to OSM.

The two ideas I outlined below are just sketches for how we could make use of that derived data to save time and effort, though there are surely many more possibilities.


On Wed, Apr 29, 2015 at 3:36 PM Pierre Béland <pierzenh at yahoo.fr<mailto:pierzenh at yahoo.fr>> wrote:
Sorry after 5 non stop days dont have time to digest everything.

We surely have to coordinate with others and not repeat what they have done. This imagery would be useful for OSM response to have general evaluations of villages + roads. We can also provide access to TMS images to hums and Nepal gov.

Monitoring is uneasy with all the imagery requests, imagery available from various sources.

It would help that the image monitoring uMap be completed with all the tasks and available imageries.

Do we have easy access to these imagery used for evaluation. Do we have tms link to easily use. Could the info be added to the uMap?



De : William Morris <wboykinm at geosprocket.com<mailto:wboykinm at geosprocket.com>>
À : "hot at openstreetmap.org<mailto:hot at openstreetmap.org>" <hot at openstreetmap.org<mailto:hot at openstreetmap.org>>; robert.soden at gmail.com<mailto:robert.soden at gmail.com>
Envoyé le : Mercredi 29 avril 2015 13h05
Objet : Re: [HOT] TomNod integration for the next time around

My bad - Robert pointed out that I pasted the wrong source data link in there (it was for TomNod Haiyan). Here's the Nepal data: https://t.co/zzSX2qrelM


On Tue, Apr 28, 2015 at 11:56 PM William Morris <wboykinm at geosprocket.com<mailto:wboykinm at geosprocket.com>> wrote:
Kevin Bullock of DigitalGlobe has graciously shared the results of the TomNod effort so far, licensed CC 3.0 (with a plan to specifically permit OSM use):
- map: https://geosprocket.cartodb.com/viz/bb5ed630-ee1e-11e4-8dca-0e018d66dc29/embed_map
- data: http://mapperdev.tomnod.com/results/

It seems like there might be an opportunity in future activations (or now) to use this data for:

1.) Planning and prioritization - as we move into damage assessment we're probably going to be a step behind the TomNod participants; we should be able to configure tasks that focus on areas with a high concentration of damage reports and high levels of agreement. (I can't claim origination of this idea: https://twitter.com/iheartcrowds/status/593170666662146048)
2.) Direct import - this is more dangerous, but there's an opportunity to directly link TomNod contributions to OSM. High-agreement damaged buildings seems like a possible starting place.

There are many technical considerations to be ironed out, including the realtime nature of the TomNod data. It could benefit from a streaming API of some sort, as well as some form of integration in the task manager. In the meantime there's this somewhat-awkward url for overlays:


Hopefully this is an opportunity to minimize effort duplication and leverage another motivated group of contributors.

-Bill Morris

HOT mailing list
HOT at openstreetmap.org<mailto:HOT at openstreetmap.org>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20150429/237a9de4/attachment-0001.html>

More information about the HOT mailing list