Citizen Science

During extreme events people in charge of rescuing activities often face a severe lack of data which make difficult to coordinate the rescue operations. As a result, it is difficult to intervene where and when the interventions are needed.

At the same time citizens involved in the disaster spread a huge amount of information on Social Media or through SMS and in the recent years more and more attention has been given to them. Even the United Nations has started to look at this stream of crowdsourced data and the United Nation Office for the Coordination of Humanitarian Affairs has created an “Hashtag Standards for Emergencies”. However, there are also other data scattered around the net which do not have a specific hashtag but that can be nevertheless important during an emergency.

In addition, there are also challenging difficulties in managing this information since they are not structured and often too specific or not geo-localised.

The aim of this project is to try to access and better manage data coming from twitter. Twitter has been chosen because in tweets hashtags are commonly used and because the maximum amount of words is limited. This last factor enhances the probability that the information contained in a tweet (which has been selected according to some key-words) could be relevant.

In order to achieve this purpose, we want to combine two different approaches: one that is top-down and one that is bottom-up. For the former we will simply define one or more hashtags to be used in an emergency and we will retrieve the tweets containing those hashtags. Instead for the latter we will try to access to all the data related to the event that fall out the previous category by retrieving data according to some key words that citizens may use. In the second step of the process we will investigate the possibility to realize a machine learning system which will automatically select the relevant material and classify it in some pre-defined categories.

One possible city in which the project may be developed is Dar Es Salaam in which floods are a serious problem.

Land Use and Land Cover maps (LULCM) are crucial for a wide range of applications, e.g. for modelling the climate and biochemistry of the Earth. These maps are usually created through the classification of satellite imagery and then validated using reference data. Even though the availability of LULCM is increasing, the process to generate and regularly update them is typically long, costly and time consuming. As a result, available LULCM are often insufficient to describe rapidly-changing environments. In addition, many of them have a level of detail and spatial coverage that is not adequate for many applications.

To address these issues, this research evaluates the potential of exploiting citizen-contributed data, particularly the OpenStreetMap (OSM) project, for generating LULCM. OSM is the most popular example of Volunteered Geographic Information (VGI). Founded in 2004, it currently has 2.5 million contributors, who have collectively made it the largest, most diverse, most complete and most up-to-date geospatial database of the world. OSM is a rich collection of vector data, which in many cases, includes geographical objects that are not traditionally available in other products (e.g. in authoritative maps). The continuous updating of the database as well as the open license under which it is available, make OSM a suitable data source to derive LULCM. Despite these advantages, there are problematic issues with OSM data, including uneven spatial coverage, and geometrical and semantic inconsistencies, which present challenges when using the data in the context of LULC mapping. For this reason, an automated procedure has been developed that converts all OSM objects available in a given area to a LULCM having the same nomenclature as two well-known products, i.e. the Urban Atlas (UA) and CORINE Land Cover (CLC). The procedure couples spatial analysis with a set of decision rules, which were defined to solve the inconsistencies generated by the overlapping of map features assigned to different LULC classes. It can also process many different relevant tags found in these regions. The procedure has been automated as a Python script and makes use of open source technologies like GRASS GIS and GDAL/OGR. The output LULCM can be then compared with the corresponding UA and CLC datasets and the agreement can be computed. The procedure, which was tested in regions of Paris, Milan and London, showed satisfactory results, which are very promising for the large-scale applicability of the algorithm. The oral presentation will explain the rationale of the procedure, present the results for the three cities, discuss the benefits and limitations of such an approach and outline future work the authors plan to do. This will include, among others, the comparison of LULCM obtained from OSM data to classifications extracted from Sentinel 2 data, and the implementation of a Web service to make the procedure available on the Web and for any user-selected area.

The authors would like to acknowledge the support and contribution of EU COST Actions TD1202 “Mapping and Citizen Sensor” (http://www.citizen-sensor-cost.eu) and IC1203 “ENERGIC” (http://vgibox.eu).

The Australian Geoscience Data Cube Programme (AGDC) envisions a Digital Earth, composed of observations of the Earth’s oceans, surface and subsurface taken through space and time stored in a high performance computing environment.The AGDC will allow governments, scientists and the public to monitor, analyse, and project the state of the Earth.The AGDC will also realise the full value of large Earth observation datasets by allowing rapid and repeatable continental-scale analyses of Earth properties through time and space.

Remote sensing and community-based approaches for forest monitoring have been growing noticeably in recent years.  However, integrated platforms for tropical forest monitoring are still scarce. This study describes an integrated community based tropical forest monitoring system which combines emerging technologies, remote sensing and community-based observation in support of REDD+ monitoring, reporting and verification (MRV) implementation. The development of this research is driven by three specific research questions: 1) how can information and communication technologies (ICT) support the automation of the community data collection process for monitoring forest change activities using modern handheld devices? 2) what is the accuracy and compatibility of community collected data compared to other data (e.g. optical remote sensing and expert field measurements) for quantifying forest changes? and 3) what is a suitable design for an interactive remote sensing and community-based near real-time forest change monitoring system and how can such a system be operationalized? The system is developed using open source technologies and has been implemented together with  local experts in UNESCO Kafa Biosphere Reserve, Ethiopia. The result shows that the system has been able to empower local community, public institutions and civil society with the information which lead to better manage and conserve forest landscapes. The methods presented in this research are applicable to a broader geographic scope. Hence, this study emphasizes that the policies and incentives should be implemented to empower communities and to create institutional frameworks for community-based forest monitoring in the tropics.

Citizen science has been practised for a long time but recently it has
been greatly revived thanks to the spreading ICT technologies and
location based services. Besides its recent fortune stems from cultural
and political changes which requires more participation and new ways of
building interaction. CS is deeply connected with the data revolution
and the information economy. In this talk I will try to address the
advantages and criticalities of CS application in the fields of social
and enviromental studies.