The E2mC Project
When a natural disaster happens, Copernicus EMS builds rapid maps based on geospatial information. The E2mC Project integrates social media analysis, crowdsourcing and machine learning. The E2mC Project has the potential to improve the timeliness and accuracy of the maps, particularly in the first hours following an event.
Crowd4ems.org is The E2mC Project’s crowdsourcing platform. Its goal is to validate and improve the information extracted by the algorithm on social media.
As a head of UX at Scifabric, I worked closely with the Scifabric’s devs team. I collected information from the Université de Genève, e-geos and other partners. Université de Genève provides the final content and communication tone.
I ideated Crowd4ems.org platform from scratch. I led the platform conceptualisation, stakeholders interviews and user validation. I led IA and prototyping.
I had several stakeholder’s interviews to obtain a shared view of the project, its goals, users and usage. To understand the project's requirements I reviewed the extensive documentation created by several project’s partners. We analysed in detail the system workflow and I defined the Information Architecture of the platform.
Let’s see an example of how it works. Imagine that an earthquake just happened. The algorithm has been activated and it starts to crawl images on Twitter. In parallel, Carlos has witnessed the earthquake. He takes a photo of a destroyed road and he shares it on Twitter. The algorithm is able to extract Carlos’ photo and millions of other images potentially useful to map the affected area. This content is updated to Crowd4ems.org. Volunteers are asked whether Carlos’ image is relevant. The image is classified as relevant and geolocated by the crowd community. The Copernicus EMS professionals use this information to improve the timeliness and accuracy of the maps provided to civil protection organisations.
We organised some activities to validate the first release. I was in charge of surveys and I run task-based usability test. We learned that while the platform felt easy to use, the purpose was unclear.
Navigation though the site lead participants to error several times. We reviewed the main navigation and screen flow to make it more intuitive.
The most complicated part was the Geolocation task. The goal wasn’t clear. We needed to provide guidance. Key improvements for next iteration were to make the interface more conversational, providing contextual help and ensuring accurate feedback.
Crowd4ems.org Beta is live since summer 2018. The E2mC Project is gathering feedback from practitioners, researchers and technology providers. A new iteration is planned by spring 2019.