FACILITATING DATA VISUALIZATION AND EXTRACTION: A NEW PLATFORM FOR THE UPPSALA CONFLICT DATA PROGRAM (IN14-0995:1)


PROJECT AIMS AND POTENTIAL CHANGES TO THE PURPOSE DURING THE PROJECT PERIOD

The purpose of the project was to build a new platform - a new website - for all the data on organized violence in the Uppsala Conflict Data Program. The UCDP data had become so vast and complex that we needed to find a solution for a platform that could meet the needs of a variety of users to easily access data. The UCDP currently holds a total of over 200 000 observations of violence, various units of analysis and large numbers of variables.

The project has by and large followed the original plan, with the components described in the original application being implemented and delivered.

The design and integration of individual components, as well as the manner in which they operate, were partially adjusted based on multiple rounds of discussions with various users of UCDP services, mainly through the European Network for Conflict Research (ENCoRe) framework, an EU-COST sponsored network, gathering all leading data-oriented research institutes in Europe.

One of the most important alterations done to the project, based on user input and user testing, was based in the realisation that different usage needs (sometimes by the same users) require different toolkits and have different usage patterns. As such, we designed two separate interface "usage flows". The first one is based on machine-to-machine communication in the form of a full-fledged, state-of-the art, public API. The second one is more exploratory, focusing on a mix of statistical data, graphics, geographic information and textual information. Both "flows" make use of the same underlying data, and can, of course, be mixed in usage.

In sum, the purpose of the project has been followed consistently throughout the course of the project period and no changes to the overall purpose were made.


PROJECT RESULTS AND DISCUSSION

The newly built infrastructure, which became operational as of April 2016, (http://ucdp.uu.se/) is now the core of UCDPs interaction with the research community, our main user base. Some components remain to be released, including three out of four API components scheduled for release in July 2017, but the interface is nevertheless fully functional for exploration. This means that all UCDP data is available to the public. Large amounts of data operating on different levels of analysis, which could previously not be easily overviewed, is now all accessible from one source.

As mentioned, the interface is built around two separate "usage flows". The first one is based on machine-to-machine communication in the form of a full-fledged, state-of-the art, public API, allowing fast access, subsetting and filtering of large amounts of data in a guaranteed consistent manner. The second one is more exploratory, focusing on a mix of statistical data, graphics, geographic information and textual information. The latter allows for exploration of both hierarchical and same-level relationships in the data allowing the user to explore and make use of the data piece by piece, without being overwhelmed. Both "flows" make use of the same underlying data and can be mixed. This allows for flexible and unrestricted usage of the data for a variety of user types and purposes.

Prior to the April 2016 launch, the new interface was shown at the UCDP exhibition booth at the 57 Annual Convention of the International Studies Association (ISA), 16-19 March 2016. The total number of visitors at the convention approached 6,000 and the UCDP booth solicited visits from significant numbers of interested researchers who viewed and tested the interface. We have also presented the interface at various other academic conferences and workshops within and outside Sweden, as well as to other audiences, including high schools. In the future, the new website will be shown at recurring ISA conventions as well as all other conferences and workshops involving UCDP researchers and staff.

Concerning reception and impact, traffic to the UCDP website has more than doubled since the release of the new interface. Download sessions have also nearly doubled. We expect much more impact after the completion of the project, judging by experiences from releases of similar infrastructures from related research projects, such as the PRIO-GRID at Peace Research Institute Oslo (PRIO). The response from users has been very positive in regards to improved usability and access to the data, as well as to visual aspects of the interface.

The project has also resulted in increased collaboration with other data infrastructures. A more consolidated relationship has developed with the International Conflict Research group at ETH Zürich, which hosts a data portal (GROWup) developed with EU funding and which contains related data in the field of peace and conflict research, as well as selected UCDP data. In particular, this collaboration concerns the person responsible for developing their infrastructure, Luc Girardin. The collaboration ensures full compatibility with other relevant data and has the potential of resulting in one or more peer-reviewed publications in either computer science or conflict research related fields. Collaboration concerning both infrastructure and research has also been initiated with 'Varieties of Democracy' (V-Dem), Department of Political Science, University of Gothenburg, and the 'Quality of Government Institute' (QoG), University of Gothenburg, two other major data infrastructures in the social sciences in Sweden.


UNFORESEEN TECHNICAL AND METHODOLOGICAL PROBLEMS

One challenge was to determine how the new interface would function and what needs it would cater to. The question was if it would be a user centered platform, or a data centered one. Initially, formal user tests were conducted to try to establish needs based on current users' behavior. It soon became clear that, given the methodological challenges posed by such advanced, cutting edge scientific requirements, standard user-centric enquiry tools used in human-computer interaction design would simply not work, as users themselves were not certain in their requirements. Instead, based on in-depth interviews, we decided that the way forward was rather to put the available data in focus and work with possible solutions given the amount and complexity of data at hand, taking into account and integrating input from users throughout the process.

Another challenge was adapting the data and the platform to work with the varying needs of researchers, which demanded everything from open machine-to-machine communication to mobile accessibility. To solve this issue we have built multiple access channels that cater for the technology each user is using to access the platform (i.e. a separate API for direct download of data, a responsive web platform for visitors using a mobile device etc.).

A third challenge was purely technical in nature and had to do with server space availability and cost at the university level, which was severely limited. We solved this issue by choosing technologies that fit the server choices made by the university administration, allowing maximum maintainability in the long term, rather than the optimal technologies for the project itself, which would have allowed for faster development and deployment, but which would not have been feasible financially.


INTEGRATION INTO ORGANIZATION AND WAYS TO TAKE THE WORK FORWARD

The interface resulting from the project is fully integrated in the organization and in operation since its launch in April 2016. All current and future work is therefore taking place within this new environment. The technical maintenance of the interface will be taken care of with the help of our permanent IT staff, funded partly by the Department, partly by a permanent yearly grant Uppsala University. We estimate that we will be able to maintain and make basic upgrades of the infrastructure without additional funding in the coming 10-year period.

The newly built infrastructure is now the core of UCDPs interaction with the research community, our main user base, a role which we envision will only deepen with time. As such, we believe the interface will require not only technical maintenance to keep it functioning but also constant improvement work in order to keep up with UCDPs offering of data, projects and future collaborations. One such example of future research initiatives that the infrastructure has made feasible and which will, in turn, require the infrastructure to make future expansions, is the 'Violence Early-Warning System' (ViEWS), part of an ERC Advanced Grant held by Professor Håvard Hegre at the Department of Peace and Conflict Research, Uppsala University (see: http://www.uu.se/en/media/news/article/?id=6434&typ=artikel&area=2&lang=en). This project focuses on forecasting of armed conflict and it employs UCDP data as one of its main data sources. The ERC grant contains some funding for upgrading and tailoring the UCDP system in line with the particular needs of ViEWS. In general, funding for this type of upgrades - which are separate from normal technical maintenance and basic upgrades - will have to be secured from within such projects, or applied for separately.


NEW RESEARCH QUESTIONS GENERATED BY THE PROJECT

The project is geared to providing easy access to a data infrastructure containing large amounts of diverse data on organized violence. In doing so, it reduces limitations for which research questions can be asked, or which may have been hindered by the costs associated with data handling. The data, in all its complexity - with a multitude of variables and levels of analysis - is accessible for exploration as well as for download. With the machine-to machine connection (API), there are no longer any limits for researchers to merge with other data, filter and make use of all UCDP data.


WEBSITE LINK

http://ucdp.uu.se