Keywords

1 Searching for Political Imagination

In Northern Ireland, during the conflict and as the peace process emerged I worked, with others, to support human rights issues being addressed in the agreement. As a ‘multi-hatter’ I was also involved in a ‘Women into Politics’ programme in Northern Ireland, aimed at supporting women in all political parties to engage in the peace process, and to form a women’s party if they wanted (they did, and did!). This programme operated in women’s centres across the country.

With the arrival of the peace process, we wondered how the peace process should and could improve human rights, equality and inclusion. Many of the human rights issues had underpinned the conflict, or become part of how it was waged. The conflict lasted 30 years—at that point most of my life. We had a series of suggested reforms even during the conflict. But what could and should we expect to change ‘all at once’ if there was a peace process? How could human rights issues be leveraged into a mediation process that would assume a central focus on sovereignty issues, and what form of drafting would capture commitments in ways that could support effective implementation?

Instinctively I, and others, looked to see what people were doing in other peace processes. How had issues like police reform, removal of emergency law, or release of prisoners been addressed? When and how did human rights issues enter processes? What types of argument were persuasive? How had mediation design enabled those arguments to be made? When and how had women influenced a peace process? What had they asked for? How might that experience benefit women in Northern Ireland?

2 Emergent Peace Analytics

I began to look for information about other peace processes. Often Northern Irish scholars and activists looked to countries with historic or solidarity connections for comparison, such as to the Israel-Palestine conflict, or apartheid South Africa and to a lesser extent to Sri Lanka. Their processes often did not ‘fit’. I looked to see if there were any other peace processes with peace agreements. To my surprise, I found a lot more than I expected, many of which I was only vaguely aware.

I collected peace agreements—increasingly obsessively! I worked to understand the conflicts and peace processes they grew from. These days were—amazingly—before the internet was fully part of our lives. To get agreement texts I wrote to both armed groups and governments. I was often surprised that they sent them in quick response—many were quite proud of the agreements they had negotiated. Other times I had visited the country to assist in mediation or met someone from it, and got hold of texts. In yet other contexts, pioneering individuals had made compilations of documents that were readily available in country, although hard to find abroad.

As the internet took off, some agreements appeared on the web only to disappear. I learnt to download upon finding, and worked to verify content. Some countries over time developed detailed web resources. The Israeli Ministry of Foreign Affairs, for example, put up all the Oslo Accords between Israel and the Palestinian Liberation Organization (PLO). Many are still there now, although reorganised and harder to find.

I began to develop a definition and list of peace agreements. I tried to stay neutral as to how long the agreement had lasted, or the form the agreement took as a ‘treaty’ or a signed text, or an interim constitution, or UN Security Council Resolution (which sometimes are modes of documenting agreement, Bell, 2006), provided it had been formally agreed. I also collected agreements regardless of whether they were successful or failed. My first ‘census’ was published in 2000 as an appendix to my first book ‘Peace Agreements and Human Rights’ (Bell, 2000).

3 PA-X Peace Agreement Database

Over-time this collection resulted in the PA-X Peace Agreements Database. This is a census of all peace and transition agreements from 1990 (when the Cold War ended and the contemporary practice largely began), to current day. It comprises over 2000 agreements found in 150 peace processes. A much different more expansive resource than that I initially contemplated and embarked on (see Bell & Badanjak, 2019)!

It includes agreements at all stage of a peace process classified into:

  • ceasefire agreements

  • pre-negotiation agreements that involve ‘talks about talks’—who is going to come to the table, with what status, what agenda, and what pre-conditions met

  • partial framework agreements, that address multiple issues but do not purport to be a final agreement

  • comprehensive agreements, that purport to offer the main promises necessary to end the conflict

  • implementation agreements, reworking agreements to include new actors or address new issues

Each is searchable by name, entity (where relevant), country, region, peace process, and stage. PA-X has over 250 searchable topics with extracts of the text, covering all the critical issues in peace negotiations. It enables both quantitative and qualitative research on actors and issues included. PA-X is best understood as a number of datasets, with a range of functions.

PA-X has a fully elaborated codebook and manual, and exists as a machine searchable corpus of texts (which enables machine learning—more on that later); and search results that can be exported as pdfs, or csv files of the topics included in a spreadsheet, or a timeline.

4 Peace Process Data

PA-X is not, however, just a database of agreements. I was interested, of course, in whether and how particular issues were included and the terms in which they were drafted. However, I was also interested in ‘peace and transition processes’: how mediation processes were designed, what parties were included at what stages, and how this shaped the issues deemed relevant to resolving the conflict. PA-X is best thought of as a peace and transition process database that enables peace processes to be traced and compared over time.

Agreements provide an important window into peace mediation trajectories and outcomes. They do not capture the sum total of what is important in a peace process, but they document the public commitments that the parties to the conflict made with a view to ending the conflict, at least for a moment of time. They therefore provide an interesting documentary trail that offers a basis for comparing peace processes, including what actors get to the negotiation table, and when, and how that influences the agendas for change adopted there.

As we visualized our data in process form –we began to be able to see and to communicate the long nature of peace processes as iterative. They moved forward and backwards, with new agreements re-shaping what had come before. This I suggest changed the mediation field. It helped researchers and practitioners to think about approaching negotiations, not as leading to a final agreement, but as ongoing complex processes of change over time (see eg Paffenholz, 2021). You can explore our ‘messy’ peace process interactive visualisation to see what the forwards and backwards movement of peace processes looks like (for an account of the process of production see, Bell et al., 2022).

Peace mediation trajectories globally, mirror ongoing implementation challenges in Northern Ireland. There an Agreement had been signed and included significant human rights and equality provisions. While we had seen it as an end-point in the conflict, we were soon taken aback in how difficult it was to get parties to stick to their promises and implement. It was interesting to compare to other peace processes, and see that an agreement is often the beginning of a new phase of negotiation, and even the peace process ‘proper’, rather than its end.

There were also other motivations for creating peace process data. I wanted to create a ‘census’ of peace agreement practice with capacity to trace processes over time, and produce descriptive statistics that would also enable multiple ways of using other data to track implementation.

Finally PA-X was driven also by trying to preserve an archive. If I had wanted these documents and found them difficult to collect and parse through, then others could benefit from my work. Plus, I cared about the documents surviving for history, as artefacts themselves.

5 PA-X Extension

Over time, we have created or incorporated other ‘interoperable’ datasets to build a broader picture of peace processes. Interoperability means that the data can be used together in ways that make sense. PA-X includes a ‘gender database’, that includes a ‘deep dive’ into gender provisions in peace agreements—this is the dataset that underpins the PeaceFem App described earlier. We also created PA-X Local, including agreements that refer to small geographic areas in broader armed conflicts, such as Atem was working on in Chap. 1.

We have linked an Amnesties, Conflict and Peace Agreement (ACPA) dataset, developed by Professor Louise Mallinder of Queens University of Belfast as part of PeaceRep. This is another ‘process’ dataset—amnesties are not events as they do not arrive in a single moment but are elaborated in processes that unfold through agreements and often legislation. By working to combine PA-X amnesty data with Louise’s earlier amnesty project, and expanding it, we had a basis for a distinctive set of data and could couple data with a capacity to create a user-friendly interface for exploring it, and ensuring interoperability with the peace agreement data.

Amnesty data is important because amnesties remain controversial in peace mediation and often pose a stumbling block to agreement. Amnesty is often the price of reaching peace with armed actors, but forgiving and forgetting sells out victims and ‘accountability’ in ways that can undermine peace agreement implementation down the line. The amnesties data helps to move beyond assertions that amnesties are either all good or all bad, by illustrating the many options for their design and evolution over time. In so doing, it illustrates innovate ways to square the amnesty-accountability circle.

6 Data Interoperability

We have also worked to create interoperability with other datasets, that we either have partnered with, or have worked collaboratively in PeaceRep to develop. This includes: Constitute, a database of the world’s Constitutions, a new International IDEA Database in Constitution-Building Processes in Conflict-Affected States, and detailed perceptions surveys that have been carried out in the field by local researchers in South Sudan, Syria and more recently Ukraine, that will be integrated with PA-X through a new Peace and Transition Process Tracker.

In the area of gender, we have created data interoperability with qualitative and quantitative data on implementation of gender commitments, to create the PeaceFem App as described in Chap. 7. We have also worked to create interoperability and a relationship between our agreements data and UCDP, Correlates of War, and ACLED conflict data. Again, these efforts to inform the development of the Peace and Transition Process Tracker described below.

7 Research and Development for Peace Analytics

Emergent peace analytics illustrate how research and development is driven in the field by collaboration between institutions involved in similar enterprises. Over time, others also became engaged in similar peace agreement collection efforts to our own, and other peace agreements collections emerged, for different purposes, with different parameters and ways of categorising and organization. USIP provided an early Peace Agreement Digital Collection, which survives as a ‘library’ in which some things are deposited and not others, around the same time as my census. UN Peacemaker—designed for mediators, started to publish agreements in the early 2000s, as part of a wider set of resources to support mediation. UCDP started collecting peace agreements that related to the conflicts and armed actors that it includes in its conflict data. This led to a collection that by 2021 had reached 375 agreements (Davies et al., 2022). The Kroc Institute for International Peace Studies, Peace Accord Matrix (PAM) at the University of Notre Dame, created data focused on the implementation of 34 ‘comprehensive agreements’ over a ten year period. ETH Zurich, has recently released a ceasefire dataset, of lateral, verbal, written and non-​implemented ceasefires from 1989 to 2021.

As these projects emerged, we worked to support these projects, sharing data and methodology, learning regarding modalities of creating interfaces, and often swapping agreement texts. In turn we benefited from how they developed their data, and reciprocal collaboration as their projects developed. Collaboration in the early days helped to bridge moments when the work was experimental and not always understood or supported. I remember once offering Stina Högbladh at UCDP a peace agreement she hadn’t been able to find, and she said ‘you know this is a little Christmas present for me’!

All of these projects have had close relationships with the mediation field of practice. It is worth noting that, like the deaths in conflict data described last chapter, each agreement dataset remains different in its purpose and scope, and therefore the scope of the collection, the classification and topic modelling approach are all different. Collectively the agreement collections have developed a ‘state of the art’ in how peace processes and agreements are traced and understood.

Collectively this work has built a foundation for data analytics in the peace building field, although none of us would have put it that way. Let me call it ‘peace analytics’.

8 ‘Peace Analytics’

Peace Analytics is a third emergent area where there is ongoing digital innovation and capacity to produce impacts of scale across conflicts, in how peacebuilding is understood and supported. See the box below.

Data analytics: The science of using raw data and computational analysis for use of discovery, communication and interpretation of patterns in the data, so as to inform about how real-world phenomenon operate in practice.

Peace analytics. This use of data pertaining to peace and conflict related issues, for use of discovery, communication and interpretation of patterns in the data, so as to inform conflict resolution and peacebuilding practice.

Data. ‘Organised information’ whether quantitative (numeric), qualitative, pictorial or text.

We have seen examples of peace analytics and relevant data in many examples in this book. For example, CEWS are data-driven, and increasingly integrate quantitative data with other forms of data. While CEWS focus on predicting and responding to conflict, peace analytics aim to support and resource peace and transition processes, although tasks and some of the data, intersect.

Peace analytics as it develops has some capacity to plug already-mentioned gaps, such as mapping peace or dialogue or civicness, and perhaps even predict its onset. However, challenges remain, as illustrated by our experience.

9 Challenges of Data Development

9.1 Software Choices

A key obstacle to developing a new dataset is finding the right software to use to produce it. The software is vital to how simple or hard the ‘back end’ of data input will be, and how user-friendly the front end will be. My earlier attempts at a database had been frustrated by using ready-made database programmes that tend to be the software installed on University computers, that did not quite ‘fit’ what we were trying design. They soon became difficult for people working with them to use in part because ‘workarounds’ to make them fit required training on systems that were not intuitive. I also very soon wanted to customise the data in ways existing software did not easily allow. My collection had massively grown to well over 1000 documents, and it was clear that direct topic-based access to the language of texts was very useful. As time progressed it felt much simpler to design a dedicated interface for coding agreements that staff could be more easily use and that would offer capacity to alter data structures or change one’s mind over coding.

9.2 Staffing and Environment

A first breakthrough was finding the right team in University of Edinburgh’s Information Services, to deliver a specially designed database with dedicated back-end and front end interfaces.

Soon, I needed to think about research staff skills. Both how to get the right skills, and also how to make laborious tasks more doable. I expanded my team to include a political scientist who was highly expert in quantitative analysis, Dr. Sanja Badanjak. This was transformative of the nature, ambition, scale and professionalism of the project. Sanja brought her own vision of what the database could be, came with a strong understanding of conflict data—and its constraints—and introduced me to terms such as ‘data interoperability’. She created a system for relating country codes that link the data to nearly all the relevant conflict data (and wider), by using ISO country codes and the Gleditsch-Ward List of Independent States, that can account for things that ISO, being an official account of countries cannot (e.g., countries that no longer exist). This development meant that we could examine not just what was in peace agreements, but their implementation as measured by the effect on deaths in conflict, triangulated across three datasets UCDP, ACLED, and Correlates of War (which deals with civil wars and interstate wars). As noted earlier, these all have different thresholds of deaths, and have different methodology, which again need to be understood before using them.

Sanja’s work meant that we could connect deaths in conflict to our data, and develop ways of standardise encoding data, and align with these datasets. She also brought a much more robust methodology to coding agreements and elaborating a code book.

We also continued to learn from other projects. For example, Constitute, the constitutions database had developed strong methodologies for ascribing text to topic. These included, an automated ‘question-and-answer’ function for enumerators of topic content to inquire about ‘borderline’ or unclear cases. A central decision-maker would respond with the decision recorded into an ongoing repository to inform consistent enumeration over time. We incorporated this design-feature into our database backend.

An amazing team of researchers also came on board, who engaged and co-developed how we defined and documented topics. The staff team developed topic definitions collaboratively, all suggesting modifications that would evolve a form of ‘good practice’. At each point we would often need to modify our ‘back end’ and ‘front end’ interfaces for coding and enabling searching of data. Out of this process evolved quantitative innovations such as ‘weighting’ provisions quantitatively depending on whether they were purely rhetorical or made a clear commitment to change.

As our work has developed we have had to add skills and components: employing data engineers; and Visualizers. We have had to think how to architect and engineer our data into a more sustainable computing environment—for us using the Edinburgh Parallel Computing Centre (EPCC) supported by a new Edinburgh International Data Facility both with ‘big data’ capacity (EIDF). We have also tried to think about how we make what we have learned and done open source and reusable, by creating a curated data hub.

This has all involved also constant ‘re-training’ and mutual learning, as we grapple with ‘large-scale computing’, CKAN platforms for curating data, data pipelines, data marts that coordinate different types of data in one database, and different ways in which databases can be created and maintained. As well as ethics, data protection, and security.

10 End-Users of Peace Analytics

We have always designed our peace process data with four main communities of end-users in mind, and all are active users of PA-X. We work in collaboration with these end-users in multiple ways, and many of the ideas of how to develop the data come from their articulation of needs and concerns.

Armed actors in conflict. These include both state security forces and non-state armed combatants, who seek to move to peace and want to know what a process looks like. They can often come from a position of little to no expert advice on peace process design because they are considering ending it for the first time. Providing a resource that armed actors can use is controversial. But if conflicts are not to continue forever, then they need to be brought to an end.

International organizations and mediators who are involved in negotiating conflicts. The idea of PA-X was that it would provide ‘Peace Agreement Access’ for mediators, to be able to draw on topic-based resources that they were encountering in negotiations, and also be able to look easily at regional or comparative examples.

Organizations focused on ‘inclusion’ in peace processes, in particular the inclusion of women. These can often be international non-governmental organizations—human rights groups, or peacebuilding organizations, who seek to try to ensure that people additional to those doing the fighting, get a chance to shape the peace agreement and the country’s future.

Civil society actors in-country. These actors often seek to understand how they might shape the peace agreement, or work to secure its effective implementation.

Researchers. PA-X now supports most of the quantitative research on peace processes, as it provides a transparent and reliable basis of counting and quantifying peace agreement commitments over the duration of a peace process. This means that the research it supports now goes well beyond our own team. PA-X has enabled a new research capacity.

11 What Does ‘Peace Analytics’ Help Us to Do?

What use is this type of data—what type of support to peace processes does it enable?

Evidence-based innovation in peace process design and drafting. Our database is used frequently by mediators—of all different types—to understand how different issues in peace agreements are drafted, what is possible at what stage in a process, and what is at stake in the drafting. We have a set of ‘Peace Agreements and…’ publications that in part originated as response to mediator requests to understand how particular issues were dealt with in other contexts. Local peace mediators, and representatives of some of the protagonists of the conflict, also use the database, and call us up from the field. The database is not a set of drafts for cut-and-paste; it is more as being a ‘tool for political imagination’ illustrating how persistent issues in conflicts are resolved, in techniques that can be creatively repurposed and reused to work in a different context.

Development and monitoring of new norms addressing peace process inclusion. Since around 2000 the UN Security Council and international courts have increasingly produced new norms or ‘laws’ as to how peace agreements should deal with certain things. For example, UN Security Council Resolution 1325 of 2000, provided that women should be included in peace negotiations, and that peace agreements should reflect gender concerns. Early data helped provide the evidence base that issues relating to women were seldom included in peace agreements, and since that time the data has been used by the UN Secretary General to monitor implementation in annual reports.

What does PA-X tell us? Has anything changed? Yes! Sort of. Women and gendered issues such as sexual violence, are more likely to be included in comprehensive peace agreements since that norm was passed, particularly where the UN has been involved in the process (Bell & McNicholls, 2019). However, women are rarely present in pre-negotiation and ceasefire talks, or at the implementation negotiations, where pathways are set, or comprehensive peace agreements renegotiated. In Colombia, for example, significant gender elements were watered down in negotiations that took place after the peace agreement with the Revolutionary Armed Forces of Colombia (FARC) of 2016 failed to be passed in a referendum. This included negotiations directly between the FARC and Church leaders, as well as with the Colombian Government, that pushed back on gender rights, including some of the world’s only peace agreement provisions addressing LGBTQ communities.

The UN Secretary General, who reviews the women, peace and security developments each year, responded to our data and analysis by recommending that new efforts take place to include women at all stages of a peace process (UN, 2015), and subsequent UN Security Council Resolutions have recommended that efforts for inclusion at early and late stages of peace processes take place (see eg., UN SC Resolutions 2242 (2015), and 2493 (2018)).

Supporting mediators, civil society, donors and international organizations, to understand that peace processes are dynamic and interactive and need long-term mediation support, rather than a once-off intervention burst at one moment. Our messy peace process visualization, and the data that underlies it has helped create a broader vista amongst international mediators, donor states and international organizations, of what a peace process is. It has shown the multiple iterations of peace that develop in multiple agreements over time. The peace agreement data provides important knowledge of context for people who often arrive in-field on short notice. But it also in a sense illustrates that a comprehensive agreement has had a ‘before’, and will have to have an ‘after’—even if international actors are no longer present. Seeing things from a longer time perspective changes the ways in which people approach negotiation in the moment.

Understanding new dynamics and trends in conflict and peace, and supporting those involved in peace processes to adapt. More recently, our work has provided an evidence-base for understanding changes in patterns peace processes, and who mediate, that has helped paint a picture of what I have called the double disruption of peace processes that has come from shifting conflict patterns. This is now of central concern to a range of actors, and is underpinned by the clear evidentiary basis of our data and analysis, and resonates with those in the field. The data illustrates quantitatively, things that I described in Chap. 4, such as:

  • comprehensive agreements are on the wane, and the UN is less involved in mediation over time

  • interim transitional arrangements which contemplate social processes of reform are a prevailing modality for ending conflict, but go wrong all over the place for similar sorts of reasons

  • we are now in an era of geopolitical competition over ‘who mediates’ in peace processes, which is part of a struggle to shape global order as codified in international legal norms

  • many countries experience ‘iterative transitions’ against a background of fragmented conflict that is simply not amendable to a ‘big elite deal, to establish a political settlement’ that once characterized peace processes

This information has impacted on how mediators and international organization view contemporary conflict resolution challenges, and respond to them. It has also radically re-shaped our research agenda.

It is worth pointing out that all of the peace analytics efforts, have made similar tyes of contribution in different ways, related to their design. One of the most interesting is perhaps that of the Peace Accord Matrix (PAM). While developed across conflicts, the PAM team had members who were very trusted in the Colombian process. PAM was written into the Colombian peace agreement as a methodology that could be employed to monitor and support the peace agreement’s implementation. The monitoring of whether agreement promises were kept, has been vital to the implementation process (see Colombia Barometer Initiative).

12 Where Next?

As we began to see in Chap. 10 on CEWS, there is an increasingly long list of data that can be understood as useful (or part of) to ‘peace analytics’. As Richard Caplan points out, peace has no settled content, so we cannot easily measure whether it has been achieved (2021). Instead, he suggests, we have to consider what we might be a ‘peace impact’ and see if there is a way to measure that.

There are, however, data gaps. Relevant data does not always exist for conflict countries, or is not always reliable across case studies, or does not cover a long enough time period to capture a peace agreement implementation period. Data-matching is needed, and remains difficult. Nevertheless an expanding array of data for measuring peace impacts exists (see box below).

Composite data sets relevant to peace (that crunch many of the datasets below). For example, Global Peace Index.

‘Institutional peace-related event’ datasets. For example, NELDA Elections dataset, and COLPUS

Political indicators. For example, V-DEM Varieties of Democracy Data base.

Conflict and Security Indicators. For example ACLED, UCDP and Correlates of War, and SIPRI on issues like military expenditure.

Economic Indicators. For example, The Government Revenue Dataset, and Global Food Prices

The new disruption of peacebuilding efforts, and the unravelling of the global order that supported them, has re-shaped where peace analytics is beginning to go, and might go—including for ourselves.

13 Peace Analytics and Double Disruption

Concern to address changing conflict and peace dynamics has led us to work to develop a Peace and Transition Process Tracker—PA-X Tracker, to support innovation in how peace processes are designed, constructed, and implemented in a changed conflict context.

PA-X Tracker attempts to build a better picture of the types of new assembling and disassembling of actors for peace and for war unfolds, in ways that could support adaptive peacebuilding and the ongoing construction of ‘multi-level’ peace processes that appear needed for multi-level conflict systems. We want to do this in ways that are sustainable over time.

There is a tube station in the London Underground where when the train pulls in there is a dangerous gap between the train floor and the platform and famously a recorded voice warns ‘Please mind the gap’. Covid dashboards illustrated the usefulness of minding gaps between areas with Covid infections, and where hospital beds, quantities of PVC masks and gloves, etc., were located.

Our project similarly aims to use Peace Analytics to suggest where to ‘mind the gap’ between peace agreement commitment and implementation. This could be the gap between promises made and actions unfolding, or between national and local perceptions of peace. We are therefore focusing on how to point to where processes are going well, and where gaps are opening up, that perhaps peacebuilding projects, including new forms of mediation, could bridge.

However, there are considerable issues that have had to be wrangled to take peace analytics in this direction, they include:

  • Agreeing and defining what peace or ‘peace impacts’ are

  • Finding appropriate data that is reliable and consistent in conflict settings, that might measure them

  • Finding ways to bring different types of data together intelligently

  • Localising data through new localised data collection, or finding ways to disaggregate it by area

  • Connecting data findings to analysis

Our enterprise in a sense acknowledges that there is no settled meaning of peace—it is something that is constructed (remember Ken Bush and his knife). We therefore have created a Tracker with ‘different ways of seeing’ progress, across very different types of knowing and data. The interface has the following components, in this initial iteration at least.

A country timeline of formal institutional change. The timeline builds on earlier the ‘process-oriented’ data interoperability we set up to bring together automated timelines of peace agreements across all stages of the process, constitutions, elections, amnesties, and coups. We have capacity to add other events to customise timelines more contextually, but the event data provides for automated update. The timeline enables a more comprehensive political process tracking over time.

Peace agreement topic analysis linked to quantitative implementation data. A dashboard links agreements to a range of ‘peace impact’ indicators. We think the need in the field is not ‘more’ data, but ‘less’ with knowledge of interoperability and data limitations built into what is chosen. So we include only customised bundles of data that have time-periods and geographies that cover conflicts, and that we, and others in the field, trust and already use, such as those listed above, including deaths in conflict data across datasets. In countries where we have local researcher field teams, we add localised perception data.

Peace agreement qualitative implementation data. We have developed a natural language processing (NLP) methodology based on semantic similarity (see Gardner, 2023), to connect agreement text to where implementation of its key commitment are monitored in reports, to produce ‘extracted’ implementation data of a narrative type. For this, we need accurate implementation reports, which for some countries exist and for others do not, and we will have to source good quality consistent narrative information. The NLP tool is reusable for other tasks such as comparing how topics are treated in different drafts of agreements, topic modelling consultation responses in an automated way, and comparing peace agreement and constitutional content.

Actor Networks. Showing who has signed agreements, what third parties were involved, connecting this data to group dynamics within country. The work has involved using named actor entity recognition techniques to create dictionaries of armed groups (Henry, 2023). We are working to extend this type of visualization to ‘mapping peace dialogues’ and examining ‘islands of agreement’, and trace the connections between them.

We aim not for a ‘measurement’ of peace, but to paint a picture of—what is going well, what is going badly, and where indicators sit together ‘oddly’. The intention is to support deliberation by actors in the field as to: is this real or an artefact of data collection efforts; why might the picture be mixed; what does it mean, and how should it be addressed?

14 Conclusion

Like Monet’s Cathedral At Rouen series of paintings, or more obviously the wonderful Atlas of Economic Complexity that has served as inspiration, the Peace and Transition Process Tracker is designed to show things from multiple perspectives, all of which give a slightly different view. It aims to support peace process design and implementation in an era of peacebuilding disruption, by providing different ways of understanding and evaluating ‘progress’ in outcomes and processes that relate to peace. It is designed to encourage questioning and deliberation rather than to offer a straightforward statistical assessment of progress. 

Questions

  1. 1.

    What types of data do you think are useful to peacebuilders?

  2. 2.

    What would make it more usable to a range of people?

  3. 3.

    What more data of what type could be helpful?