key: cord-0434204-nufzk93v
authors: Lepuschitz, Raphael; Stoehr, Niklas
title: SeismographAPI: Visualising Temporal-Spatial Crisis Data
date: 2021-07-27
journal: nan
DOI: nan
sha: 6ea64f4a84f7a1c05c244a70db3ee783703df543
doc_id: 434204
cord_uid: nufzk93v

Effective decision-making for crisis mitigation increasingly relies on visualisation of large amounts of data. While interactive dashboards are more informative than static visualisations, their development is far more time-demanding and requires a range of technical and financial capabilities. There are few open-source libraries available, which is blocking contributions from low-resource environments and impeding rapid crisis responses. To address these limitations, we present SeismographAPI, an open-source library for visualising temporal-spatial crisis data on the country- and sub-country level in two use cases: Conflict Monitoring Map and Pandemic Monitoring Map. The library provides easy-to-use data connectors, broad functionality, clear documentation and run time-efficiency.

For mitigating large-scale crises such as armed conflicts, pandemics and natural disasters, incorporation of data in decision-making is becoming indispensable [1, 3, 7, 10, 13] . However, insights from large amounts of data remain untapped if they are not detected and communicated by means of intuitive, accurate and preferably interactive scientific visualisation [5, 8] . Particularly, the development of interactive visualisation dashboards requires a broad skill set, ranging from statistical, design and programming knowledge to domain expertise [6] . Academic environments, non-governmental and humanitarian aid organisations often lack the required resources which hinders urgently needed contributions. The demand for quick crisis responses stands in stark contrast to time-consuming, expensive development stages. SeismographAPI is an actively maintained, open-source library for the visualisation of temporal-spatial crisis data that combines plug-and-play visualisations with versatile functionality.

SeismographAPI is designed for data analysts to identify patterns in rapid prototyping. Due to its run time and memory-efficiency, it can also be deployed as a permanent visualisation tool for use by * Both authors contributed equally to this work. To motivate and demonstrate SeismographAPI , we sketch out two practical use cases that are inspired by real-world visualisation needs [2, 4, 7, 11, 13] .

Conflict Monitoring. With the help of SeismographAPI , we visualise a huge dataset comprising 20 years of conflict data on 141 countries, constructed from ACLED [9] and UCDP GED [12] data. Per country and month, our dataset features 60 socio-economic and political indicators, which are all displayed in our Conflict Monitoring Map.

Pandemic Monitoring. Our second demonstration case is the Pandemic Monitoring Map, a visualisation of COVID-19 infection numbers. The data is borrowed from Johns Hopkins University [2] .

World Map (center). The SVG Choropleth map represents the core part of SeismographAPI . It allows visualising data at the countryand subcountry-level (political subdivisions) based on the ISO-3166 and ISO-3166-2 norm. Additional information, such as country-level infection numbers, can be easily displayed on click and hover as exemplified in the Pandemic Monitoring Map. Auxiliary Information Panel (right). At the top of the auxiliary information panel, our library provides a menu allowing to interactively customise the dashboard. Users can hide information and panels, such as country names and the country list on the left hand side, zoom-in, choose a night mode and open a "help" window. To simplify the interface between analysis, report and decisionmaking, the library has built-in functionality for screen recording. Due to tight integration with Chart.js, any chart visualisation can be selected and displayed in the right-hand panel based on data suitability and information needs. For instance, the Conflict Monitoring Map displays the most important data features considered for conflict prediction as a horizontal bar chart. The Pandemic Monitoring Map relies on stacked line charts to map out infection numbers. map, rendering starts between 300ms and 800ms, document completion is done between 400ms and 2.6s and the full loading time varies from~3s to~10s. To optimise loading and usability, Seismo-graphAPI can also be initialised asynchronously with the JavaScript async/await/resolve method. After the first initialisation of the map, this enables loading data chunks on demand, which increases smoothness. This is demonstrated in the Conflict Monitoring Map, where all global conflict data (~1,1MB) is loaded at startup, but the large amount of detailed conflict data (~80KB per country,~21MB in total) is loaded asynchronously on request. Thus, SeismographAPI is able to visualise more than = 170,000 data points in the Conflict Monitoring Map in about 3 seconds or nearly = 400,000 data points in the Pandemic Monitoring Map in about 10 seconds.

Ease of Use. With an intuitive interface and simple data connectors, SeismographAPI is designed for ease of use in common visualisation tasks and workflows. Data can be loaded directly via JSON, CSV or as an HTML table. We even offer a Pandas extension to load Pandas Dataframes (as JSON) and Wikipedia tables. The library features clear readme instructions and rich documentation.

Future versions will include more data connectors, default charts, more detailed guidelines for deployment and options for switching between different data within one map. We presented Seis-mographAPI , an open-source library aimed at reducing resource constraints and easing swift data visualisation, thereby improving data-driven decision-making for humanitarian purposes.

Improving Quantitative Studies of International Conflict: A Conjecture

An interactive web-based dashboard to track COVID-19 in real time

Measuring Proximity Between Newspapers and Political Parties: The Sentiment Political Compass

Predicting Armed Conflict

Explaining the Gap: Visualizing One's Predictions Improves Recall and Comprehension of Data

Empirical Studies in Information Visualization: Seven Scenarios

Crisis Early Warning and Decision Support: Contemporary Approaches and Thoughts on Future Research

World Spatiotemporal Analytics and Mapping Project (WSTAMP): Discovering, Exploring, and Mapping Spatiotemporal Patters across the World's Largest Open Source Data Sets

Introducing ACLED-Armed Conflict Location and Event Data

Endogenous versus Exogenous Origins of Crises. In Extreme Events in Nature and Society, The Frontiers Collection

The CoRisk-Index: A data-mining approach to identify industry-specific risk assessments related to COVID-19 in real-time

Introducing the UCDP Georeferenced Event Dataset

Predicting Conflict in Space and Time