Difference between revisions of "Mapillary"

From Archiveteam
Jump to navigation Jump to search
(Change status to Online; I see no reason to mark it Endangered.)
(Mention impossibility to get raw photos, update stats on photo count and data size, and more.)
Line 11: Line 11:
Many [[OpenStreetMap]] users make use of Mapillary and Mapillary images as their source and documentation of their mapping process (a "source of truth") and as such it would be tremendously useful to be able to rely on Mapillary images being there "forever". Mapillary is a startup and as such may or may not be around forever, so having a safe backup would be a great relief to these mappers.
Many [[OpenStreetMap]] users make use of Mapillary and Mapillary images as their source and documentation of their mapping process (a "source of truth") and as such it would be tremendously useful to be able to rely on Mapillary images being there "forever". Mapillary is a startup and as such may or may not be around forever, so having a safe backup would be a great relief to these mappers.


Mapillary are very supportive of openness and freedom and it may be possible to get a lot of cooperation from them and obviate the need to do any sort of scraping.
While they keep the original photo files as uploaded by users, it's not currently possible to download them back. You can only download the processed photos: with sensitive information blurred out (faces and license plates), a watermark added on the corner, maximum 2048px width, and (apparently) 75% JPEG quality.


Some back-of-the-envelope calculations from IRC:
However, Mapillary are very supportive of openness and freedom and it may be possible to get a lot of cooperation from them, such as access to the original photos, or a dump of the processed photos without having to scrape them.


<JesseW> ris: mapillary currently claims to have 66,859,731 photos, and the downloads seem to be 2048x1536px, or about 300 KB per photo
Mapillary currently has 79 million photos. The 2048px processed photos seem to be around 300KB, which would add up to 24.3TB. Staff has indicated they are currently storing 300TB of data, but it's unclear if this is only the originals, or if it ''also'' includes the processed images in multiple resolutions.
<JesseW> Which would give a total of about 20TB

Revision as of 00:36, 22 August 2016

Mapillary
URL https://www.mapillary.com[IAWcite.todayMemWeb]
Status Online!
Archiving status Not saved yet
Archiving type Unknown
IRC channel #archiveteam-bs (on hackint)


Mapillary is a crowdsourced Google StreetView-like platform that allows users to take photos of their local area with a supplied smartphone app (or their own equipment) and will assemble the sequences in a semi-intelligent way for easy navigation and use. Photos are shared publicly under a Creative Commons license (CC-BY-SA 4.0 International license at time of writing).

Many OpenStreetMap users make use of Mapillary and Mapillary images as their source and documentation of their mapping process (a "source of truth") and as such it would be tremendously useful to be able to rely on Mapillary images being there "forever". Mapillary is a startup and as such may or may not be around forever, so having a safe backup would be a great relief to these mappers.

While they keep the original photo files as uploaded by users, it's not currently possible to download them back. You can only download the processed photos: with sensitive information blurred out (faces and license plates), a watermark added on the corner, maximum 2048px width, and (apparently) 75% JPEG quality.

However, Mapillary are very supportive of openness and freedom and it may be possible to get a lot of cooperation from them, such as access to the original photos, or a dump of the processed photos without having to scrape them.

Mapillary currently has 79 million photos. The 2048px processed photos seem to be around 300KB, which would add up to 24.3TB. Staff has indicated they are currently storing 300TB of data, but it's unclear if this is only the originals, or if it also includes the processed images in multiple resolutions.