Measuring the state of global open data for disasters and climate change adaptation

By Pierre Chrzanowski, Open Data Specialist at OpenDRI; and Grace Doherty, Geospatial Consultant at OpenDRI

A few months ago, OpenDRI released the beta version of the OpenDRI Index, a new tool to track and measure the availability of disaster risk data at country level. Today, we are pleased to publish the Global Datasets page, a preliminary list of worldwide or near worldwide coverage disaster risk datasets with an indication on whether the data is open or restricted to access and re-use.

Through the OpenDRI Index, we aim to advance the state of open data for disaster and climate risk management around the world, providing a better picture of what is available as open data and identifying essential data that is not yet available. 

In disaster risk management and more generally in the development sector, one of the trends is to rely more and more on global level datasets – that is, data with worldwide or near worldwide coverage – offering enough information and details for country, city or even district-level projects. 

For instance, there are now several projects offering world-level information on human settlements: critical data for anyone who needs to know where vulnerable populations are. 

Global datasets are often developed as a response to a lack of accurate disaster risk data in many parts of the world, but they have other advantages. A global dataset offers a standardized and very convenient way to look at and compare different territories.

What is disaster risk data? What is open?

The OpenDRI Index provides a list of high-value datasets considered essential for understanding and measuring disaster risk. Those different key datasets are classified into five risk sector categories (base data, exposure, hazard, vulnerability and risk) and include datasets such as digital elevation model, population, infrastructures or more specific disaster information such as records of hazard events or flood maps.

For each key dataset definition, users have the possibility to add one or more dataset instances and determine their level of openness according to a set of 10 open data criteria which consider technical, legal and cost issues for access, sharing and use. A dataset is considered open data where all or almost all criteria are met. 

Below is a summary of the state of global level disaster risk datasets identified by the GFDRR-Labs and OpenDRI. Please note that this is a work in progress to which you are encouraged to add your own information.

Preliminary results for global datasets

Most of the global datasets identified through this preliminary survey are available as open data or with some restriction. Out of 37 datasets tracked, not one is considered closed (access not permitted for external users), 16 have technical, legal or cost restrictions and 21 have been cleared as open data.

What restrictions are we facing? Looking more closely, the main restriction concerns datasets that are not available in an open license, are under unclear terms of use, or contain legal terms prohibiting some types of reuse. 

Recommendation: Open licensing is not an anecdotal issue. Because of legal uncertainty, users may be prevented by their organization to re-use critical disaster risk information in disaster risk management projects or, worse, during a response to disaster. Therefore, licensing should always be considered carefully by data publishers when releasing their datasets.

Another common restriction encountered is datasets not downloadable at once (in bulk), only readable through Web Map Service, or slowed down by technical restrictions to download, such as user registration. 

Recommendation: When doing a disaster risk assessment, data collection and preparation is often the most time-consuming task, more so than the risk assessment itself. Reducing time to retrieve the data is therefore critical for any disaster risk management project. Releasing datasets through a direct download should be the default option for any organization whose mission is to foster data re-use.

Lastly, no global level datasets could be identified for some of the most essential information in disaster risk management, such as building footprints or other exposure data. If you are aware of any missing key risk dataset at global level, please contribute on the website.

Improving the state of global open data for disasters and climate change adaptation

Unsurprisingly, the availability and ease of use of global datasets is better than most country level disaster risk data. After all, the primary purpose of most of these global products is to be used as a replacement for missing or restricted country level data. 

However, while they may offer a possible alternative, global data do not always meet the requirements of site-specific risk studies for which additional data collection may be needed.

This survey is a work in progress but it already offers some good insights on the state of global datasets for disaster and climate change adaptation:

• Open licensing should be considered more seriously by international organizations and other global datasets providers.
• Technical restrictions on download are not anecdotal and datasets should be available through a direct download link.
• Some of the most needed data for disaster risk management, such as building footprints are still not available at global level, something worth considering by organizations investing in such global data infrastructures.