Testing UNOSAT’s Cybermappr tool

[This is a guest blog post by Lars Bromley, Analyst at UNOSAT]

Dear SBTF,

I wanted to convey my deep appreciation for the time that you recently spent testing and offering suggestions for the Cybermappr tool under development by the United Nations Institute for Training and Research / Operational Satellite Applications Program (UNITAR/UNOSAT) and the Citizens Cyberscience Centre (CCC). Cybermappr is an experimental effort that begins to address a need by UNOSAT for converting the many photos and other media that appear during crisis periods into categorized and georeferenced data via crowdsourcing.

While its true that all media will probably one day be categorized and georereferenced at the source, UNITAR/UNOSAT felt that this future was too slow in coming. Effectively utilizing pictures and other media captured by mobile phones and similarly ubiquitous technologies in crisis response was a slow and tedious process for staff to undertake,  so Cybermappr is intended to farm that task out to the crowd. The Cybermappr trial with SBTF attempted to address two primary questions in this regard: can crowds reliably find and categorize media across the internet based on a few simple guidelines, and; can crowds then map that media to a specific location based on clues found in the media and any related text? For this experiment Cybermappr was oriented to locating and mapping photos of buildings in Libya damaged during the recent conflict.

While the SBTF volunteers assigned to this effort were not quite a crowd they did provide critical input, feedback, and guidance to the Cybermappr effort. More than 200 new photos were added and filtered, a third were ‘rejected’ as invalid, a quarter were linked to other photos as showing the same location, and only a few were actually placed on the map. More broadly, SBTF verified that media can indeed be located and categorized in rapid fashion, which is no surprise of course, but also that the georeferencing aspect is extremely hard for everyone so the Cybermappr georeferencing interface needs a lot of improvements to be effective. The georeferencing aspect is the main point of Cybermappr so we’re grateful for the input on what’s needed to make it better. Several of the specific comments we received from SBTF are currently being implemented by our developers and some of the deeper issues will be discussed at the upcoming Citizen Cyberscience Summit (http://cybersciencesummit.org/)

So, thank you all for your help we look forward to re-engaging with you once Cybermappr gets to its next phase!

Best,

Lars Bromley

Please follow and like us:
Posted in ALL POSTS, Other posts

Leave a Reply

Your email address will not be published. Required fields are marked *

*