Firstly, most of us agreed that being able to reproduce the result of queries (and potentially other transformations or processes) applied to data or subsets of the data was the hardest of the guidelines to implement.
One can deal with this by keeping archived copies of all such query and transformation results (painless to implement, but potentially devastating from a storage provisioning perspective), or one could opt to store the query and transformation instructions themselves, with a view to reproducing the query or transformation result at some point in the future.
This second option equates to always starting with base ingredients (egg yolks, lemon juice, butter, and maybe mustard or cayenne) and to store this with a recipe (in this case for Hollandaise Sauce). This option is also painless to implement, until there is a change in the underlying database schema, code, or both—in which case one will have to (potentially almost ad infinitum) maintain backward compatibility so that historical operations continue to work, or maintain working copies of all historical releases for the purpose of reproducing a query or transformation result at some point in the future. Clearly this is not very practical.
By the way, there were some excellent ideas on how to record recipes systematically: Lesley Wyborn presented work on defining an ontology whereby queries and transformations could be documented as an automated script, and Edzer Pebesma and colleagues are conceiving an algebra for spatial operations with much the same objective in mind.
This approach, of course, requires an additional consensus: at what point do we store results as a new dataset instead of executing a potentially longer and longer list of processes on original data? There must be some value to buying Hollandaise Sauce off the shelf for our Eggs Benedict—at least some of the time.
This assertion set me thinking about the process of reproducing results in the new world of data-intensive science, a world in which code and systems are increasingly distributed, reliant on external vocabularies, lookups, services, and libraries (that may be themselves referenced by persistent identifiers). None of these resources, which may have a significant outcome on the result of a process should they change, are under the control of the code running in my environment. Which brings us to Claerbout’s Principle:
"The scholarship does not only consist of theorems and proofs but also (and perhaps even more important) of data, computer code and a runtime environment which provides readers with the possibility to reproduce all tables and figures in an article."
Easier said than done. We can, of course (as we should in a world of formal systems engineering) insist on proper configuration control and versioning of all components, internal and external, but I am not convinced that the research community is ready for this level of maturity—typically reserved for moon rockets and defense procurement, with orders of magnitude in additional costs. Perhaps more importantly, the scientists writing code are not going to invest time and effort to document, version, and package their code to a standard that supports reproducibility. Hence, the code that we use to transform our data, whether we like it or not, will not automatically produce the same result at some unspecified point in the future, and much more so if it has external web-based dependencies (which, in turn, may also have external dependencies). There is some utility in packaging entire runtime environments (much in the way that one could persist the result of a query or transformation), but this does not solve the problem of external dependencies.
Which raises an interesting dilemma: in the world of linked open data, the semantic web, and open distributed processing, the state of the web at any point in time cannot be reproduced ever again—which may create significant issues for reproducible science if it uses any form of distributed code.
Not only that! As we rely more and more on processing enormous volumes of data by digital means, we will depend more and more on artificial intelligence, machine learning, and automated research. As the body of knowledge available to automated agents changes, so presumably, will their conclusions and inferences.
So...we need a new consensus on what science means in the era of data-intensive, increasingly automated science: our rules, notions, and paradigms will soon be outdated.
Fitting subject for an RDA Interest Group, I would think.
It is often said—disparagingly—that America’s culture is a consumer culture. Although it may be true that America’s consumerism is problematic, not least for the planet, the flip side is how consumer culture drives a service mentality in businesses and government. The old adage that “the customer is king” does motivate US government agencies and government-supported centers, including NASA’s Distributed Active Archive Centers (DAACs), to innovate and improve services in response to user feedback and evolving user needs.
Since 2004, NASA’s Earth Science Data and Information System (ESDIS) Project [WDS Network Member] has commissioned the CFI Group to conduct an annual customer satisfaction survey of users of Earth Observing System Data and Information System (EOSDIS) data and services available through the twelve DAACs. The American Customer Satisfaction Index (ACSI) is a uniform, cross-industry measure of satisfaction with goods and services available to US consumers, including both the private and public sectors. The ACSI represents an important source of information on user satisfaction and needs that feeds into DAAC operations and evolution. This may hold some lessons for WDS data services more broadly as they seek feedback from their users, and endeavor to expand their user bases and justify funding support.
The ACSI survey invitation is sent to anyone who has registered to download data from the NASA DAACs. In the past registration was ad hoc, and each DAAC had its own system. In early 2015, ESDIS began implementing a uniform user registration system called EarthData Login that requires that users establish a free account before they can access datasets. Accounts are associated with a given DAAC, but they allow access to data across all the DAACs. All those who register are sent invitations to fill out the ACSI survey. Response rates vary from a few percent among most DAACs, to as high as 38% for the Land Processes DAAC [WDS Regular Member] (which also has the highest number of respondents at just over 2,000).
In 2015, the overall EOSDIS ACSI was 77 out of 100, which is better then the overall government and National ACSI scores for 2015 (64 and 74, respectively), but lower than the National Weather Service (80). This score is based on users’ overall assessment of satisfaction with each data center based on expectations and comparison with an “ideal” data center. The ACSI model provided by the CFI Group also assesses specific “drivers” of user satisfaction—customer support, product search, product selection and order, product documentation, product quality, and data delivery—and their relative importance to the overall ACSI score. This allows the DAACs to identify areas where improvement is needed and should have the most impact on overall satisfaction.
The ACSI enables the EOSDIS to assess changes from year to year. For example, from 2014 to 2015 customer support went from 89 to 86, with drops in professionalism, technical knowledge, helpfulness in correcting a problem, and timeliness of response (all statistically significant). Many changes likely reflect the fact that the pool of survey respondents changes over time, as do their expectations, rather than actual drops in service provision. But for individual DAACs, declining scores in certain areas, in combination with free-text responses to open-ended questions, can help to flag issues that are in need of attention.
For example, the ACSI scores and free-text responses to open-ended questions helped our DAAC—the Socioeconomic Data and Applications Center (SEDAC) [WDS Regular Member]—in undertaking a major website overhaul in 2011. From a disparate set of pages with different designs, we created a coherent site with consistent navigation. The resulting site was evaluated very favorably by Blink UX, a user experience evaluation firm that reviewed all of the DAAC websites. Deficiencies in data documentation for selected datasets have also been pointed out by survey respondents, and we are now reviewing our guidelines for documentation to ensure that all datasets meet a minimum standard. Some users indicated difficulty in finding the latest dataset releases, so we are developing an email alert system for new data releases.
At the Alaska Satellite Facility (ASF) DAAC [WDS Regular Member], the ACSI results have been very helpful in getting a sense of how people are using ASF DAAC data and services. The free-text responses to questions regarding new data, services, search capabilities, and data formats are particularly informative. For example, one user suggested that it would be useful to have quick access to Synthetic Aperture Radar data for specific regions in the world for disaster response. A data feed was developed after the recent Nepal earthquake that notified users of any new Sentinel-1A data received at ASF DAAC for that specific area. This data feed quickly provided additional data for disaster responders and researchers studying this event. Data feeds are now available for several seismically active areas of the world that have been designated by the scientific community (i.e., Supersites).
Overall, the strong EOSDIS ACSI scores have been important in objectively demonstrating and documenting the continuing value of EOSDIS and the individual DAACs to the broad user community. The annual score is reported as one of NASA’s annual performance metrics, supporting NASA’s goal to provide results-driven management focused on optimizing value to the American public.
Although surveys can be costly, and the response rates low, WDS Members would do well to consider periodic surveys of users. We find that highly motivated users do respond and provide really useful suggestions, especially if they find that their responses actually lead to tangible changes in their user experience. While annual surveys may be more than is needed, surveys every 2–3 years could provide your data service with valuable feedback on its content and services. And of course, none of this should supplant other mechanisms for gathering user feedback, such as help desk software (e.g., UserVoice used by SEDAC or Kayako used by NASA’s EarthData), email, and telephone helplines. Through these multiple mechanisms, our user communities can help drive significant improvements in the services offered by WDS Members and the successful use of our valuable data by growing numbers of users.
A Blog post by Nico Mölg (Glaciology and Geomorphodynamics Group, WGMS)
Which glaciers are still advancing? How many are melting? Which glaciers are being monitored in your country?
A new smartphone application from the World Glacier Monitoring Service (WGMS; WDS Regular Member) shows how glaciers have evolved around the globe. It provides easy and public access to glacier observation data and photographs of more than 3700 glaciers. The wgms Glacier App—recently launched at a side-event of COP21—is based on a comprehensive research database and aims at bringing corresponding facts and figures to decision makers, to outdoor people, researchers, and anybody interested in the topic, in order to provide information and raise awareness of ongoing climatic changes.
The wgms Glacier App shows all observed glaciers on a satellite map. Basic information is provided for each glacier, including photographs and general information on size and elevation. A text search allows users to filter the glaciers by name, country, region, and measurement type. For example, one can find out which glaciers have gained or lost ice over the past decade. A compass shows the closest observed glaciers in all directions from the user’s current position, and a 'glacier card game' enables users to compare the best observed glaciers in the world and compete against the computer. In addition, graphs with observation data illustrate the glacier's development, along with information on local investigators, and detailed explanations of measurement types. WGMS wants to increase the visibility of the hundreds of glacier observers around the globe whose work documents the impact of climate change on glaciers.
Jointly developed by the WGMS and Ubique – Apps and Technology, the app is available free of charge for Android and iOS in English, German, Russian, and Spanish.
The WDS Scientific Committee (WDS-SC) requests WDS Members to maintain the quality of their data and services. Another important task for the WDS-SC is to recruit data centres from various disciplines, as many and wide as possible, to serve their data to promote science—in particular interdisciplinary science. However, based on my experiences as a researcher of Solar–Terrestrial Physics and as Director of the World Data Centre for Geomagnetism, Kyoto (WDS Regular Member), I believe that we have one more important task for promoting science: collecting and serving useful data from the 'dark long tail' of datasets.
There are a huge number of datasets—mainly obtained on a research project basis—that are not registered to active data centres, and hence are 'dark' to many of us. These datasets are typically built by small research groups for a limited period, and data quality checks are often not sufficient. Although their quality may not be good and they exist only for a limited period, such data are very important and useful if the location of observation site is highly unique, or if other observations are not available.
We know of many such 'dark long tail' datasets, and some have been sent to our data centre, but even if we find them and can ingest them, we often have difficulty to keep (or to confirm) their quality. Nevertheless, my personal opinion is that these data should also be served by WDS Members, even if they conflict with the membership requirements of WDS.
One way to compromise for the data quality and service of data from the 'dark long tail' is to register metadata that describe the observations in as much detail as possible. An example of this in practice is IUGONET (Interuniversity Upper atmosphere Global Observation NETwork), which has a common database of metadata and forms a virtual data centre of distributed databases at several institutions. This data system includes databases from the 'dark long tail', as well as large well-known databases.
The WDS-SC and WDS Member Organizations must therefore take action (and advocate) to ensure such 'dark' datasets are registered in appropriate data centres or systems with adequate metadata to make them useful. Otherwise, I have a concern that they may just be kept by each institutional repository in a way that cannot be exploited or could even be lost forever.
To improve the situation domestically, we held two workshops at Kyoto University last autumn that explored possibilities for collaboration among Japanese university libraries, informatics experts, and research scientists. University libraries in Japan are not very positive in general about functioning as repositories for scientific data. In contrast, some researchers are actively trying to develop the related technology or systems for that to happen. Moreover, a Japanese endeavour to register datasets and attach Digital Object Identifiers started last year. My hope is that these activities grow and form a stream of open data from the 'dark long tail'.
I would like to introduce a new initiative of DANS (Data Archiving and Networked Services) in the Netherlands. During the International Open Access Week last month, DANS launched, together with the Dutch publisher Brill, a new Research Data Journal for the Humanities and Social Sciences. The Research Data Journal is a digital-only, open access journal, which documents deposited datasets through the publication of data papers. The journal concentrates on the Social Sciences and the Humanities, covering history, archaeology, language and literature in particular.
Data papers are scholarly publications of medium length containing a non-technical description of a dataset and putting the data in a research context. Each paper gets a persistent identifier providing publication credits to the author.
Data papers call attention to particular research datasets, which may increase the likelihood that the datasets could be re-used or re-purposed by other researchers in the future. Additional benefits are that they are peer-reviewed, can be listed on CVs, and can accumulate citations just like traditional journal articles. This way they provide important incentives for researchers to put time and effort into preparing their datasets for public access.
The DANS Research Data Journal is an enhanced publication in more than one respect. The text is enhanced with direct links to datasets in the long-term repository. Additionally, the journal is enriched with features that contribute to greater usability of the content in terms of overview and navigation by adding background information and various forms of visualization. Where possible, data can be previewed and explored online, rather than through time-consuming downloads and offline applications. In short, an enhanced data paper provides an integrated view of data in their research context.
At DANS we hope that this initiative will stimulate researchers in the Netherlands and abroad to make their data more easily available to others.
Just returned from a useful trip to visit my collaborators working with the 'Chinese plant trait database' at the Northwestern Agricultural and Forestry University in Yangling, China. We now have information from several hundred sites across China, and this will allow us to make detailed analyses of trait–climate relationships. But trips like these remind me that data are the bitcoin of Chinese science. It requires management of complex social networks to put together a dataset this large, but there are always people outside these networks who nevertheless could contribute. And then, once the science is done, what happens to the data? There is an international database for plant trait data, but where is the trusted repository for such data in China? My young collaborators are keen to share openly with other scientists and we need to make this easier. China is one of the few countries that has a national group supporting WDS activities – guess what I am going to be talking to them about next? Sandy H.