RAMP VIS: Making Visual Analytics an Integral Part of the Technological Infrastructure for Combating COVID-19
Lead Research Organisation:
University of Oxford
Department Name: Engineering Science
Abstract
Computational modelling of the COVID-19 pandemic has been playing a significant role in the UK's effort to combat COVID-19. Across the country, there are about 100 research teams working on different models, and several dozens have provided simulation, estimation, and prediction to inform the governmental decisions in the four home nations. One noticeable oversight is the under-utilisation of visualization and visual analytics technology in supporting the scientific workflows for model development, which typically consists of a set of iterative processes, such as (a) hypothesis formulation and causality analysis; (b) model development, testing, validation, and comparison; (c) model deployment, monitoring, and improvement; and (d) scientific and public dissemination.
Because visualization is widely mistaken only for information or knowledge dissemination, the technology is commonly underused in all other stages of a modelling workflow. Ideally, modelling scientists and epidemiologists could have a quick glance of dynamic data anytime there is a need (cf. stock brokers observing stock market data), access effective overviews of spatiotemporal patterns of the disease development and control (cf., meteorologists observing satellite images, contour maps, etc.), be provided with external memorization of data to stimulate hypotheses and contemplate various decisions (cf. a general pacing around in a war room in front of many maps), and receive advice from an ensemble of analytical algorithms and visualizations about similarity, anomalies, clusters, correlation, causality, and association hidden in the data (cf. a CEO consulting specialists). Ideally, there is a visual analytics infrastructure, as a "standing capacity" (Secretary of State), that can support many modelling teams performing daily observational, analytical, and model-developmental tasks.
The proposed VA technical and knowledge infrastructures are essential for combating COVID-19 pandemic, as many epidemiologists are preparing for COVID-10 to be a threat for some time. With the recent introduction of localised control measures, it indicates an additional need for localised VA supports for many local scientists and healthcare professionals. When vaccination starts, there is a need for monitoring and modelling the effectiveness of different vaccines used in different regions. Such a nationwide need can be cost-effectively delivered by the VA technical and knowledge infrastructures.
Because visualization is widely mistaken only for information or knowledge dissemination, the technology is commonly underused in all other stages of a modelling workflow. Ideally, modelling scientists and epidemiologists could have a quick glance of dynamic data anytime there is a need (cf. stock brokers observing stock market data), access effective overviews of spatiotemporal patterns of the disease development and control (cf., meteorologists observing satellite images, contour maps, etc.), be provided with external memorization of data to stimulate hypotheses and contemplate various decisions (cf. a general pacing around in a war room in front of many maps), and receive advice from an ensemble of analytical algorithms and visualizations about similarity, anomalies, clusters, correlation, causality, and association hidden in the data (cf. a CEO consulting specialists). Ideally, there is a visual analytics infrastructure, as a "standing capacity" (Secretary of State), that can support many modelling teams performing daily observational, analytical, and model-developmental tasks.
The proposed VA technical and knowledge infrastructures are essential for combating COVID-19 pandemic, as many epidemiologists are preparing for COVID-10 to be a threat for some time. With the recent introduction of localised control measures, it indicates an additional need for localised VA supports for many local scientists and healthcare professionals. When vaccination starts, there is a need for monitoring and modelling the effectiveness of different vaccines used in different regions. Such a nationwide need can be cost-effectively delivered by the VA technical and knowledge infrastructures.
Publications
Bach B
(2023)
Dashboard Design Patterns.
in IEEE transactions on visualization and computer graphics
Dykes J
(2022)
Visualization for epidemiological modelling: challenges, solutions, reflections and recommendations.
in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Firat E
(2023)
PCP-Ed: Parallel coordinate plots for ensemble data
in Visual Informatics
Khan S
(2022)
Propagating Visual Designs to Numerous Plots and Dashboards.
in IEEE transactions on visualization and computer graphics
Khan S
(2022)
Rapid Development of a Data Visualization Service in an Emergency Response
in IEEE Transactions on Services Computing
Mitchell SN
(2022)
FAIR data pipeline: provenance-driven data management for traceable scientific workflows.
in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Rydow E
(2022)
Development and Evaluation of Two Approaches of Visual Sensitivity Analysis to Support Epidemiological Modeling
in IEEE Transactions on Visualization and Computer Graphics
Rydow E
(2023)
WITHDRAWN: RAMPVIS: A visualization and visual analytics infrastructure for COVID-19 data
in SoftwareX
Sondag M
(2022)
Visual Analytics of Contact Tracing Policy Simulations During an Emergency Response
in Computer Graphics Forum
| Description | Through the working with epidemiologists and modelling science between June 2020 and January 2021 (volunteering period) and between February 2021 and January 2022 (funded period), we have obtained the following findings: (1) It is necessary to provide many epidemiologists and modelling science and their epidemiological modelling workflows with visual analytics support through an appropriate visualization infrastructure. Such an infrastructure should ideally be pre-existed as part of the resilience infrastructure. When it is not available, it is feasible to develop such an infrastructure within a relatively short period (Khan et al. TSC, 2022). (2) While data visualization is commonly appreciated as a means of information dissemination and public engagement, it is important to make use of data visualization for supporting domain experts in performing their routine and complex scientific tasks, such as rapid data observation, data analysis, and model development. For data visualization to support complex scientific tasks, the close collaboration between visualization scientists and domain experts is essential. (3) COVID-19 presents a challenge for visualizing a huge amount of data that dynamically changes on a daily basis (Mitchell et al PhilTransA, 2022), and pushes the research agenda of automated visualization to the front. We addressed this challenge through several innovations, including the use of ontologies as knowledge bases in automation, automatic and semi-automatic agents, automated propagation of visual designs to generate numerous data visualizations (Khan et al. TVCG, 2022), and automated generation of storytelling visualizations based on algorithmic interpretation of different semantic patterns in data streams corresponding to different regions. (4) We have developed several new visual analytics tools with novel visual designs and advanced analytical algorithms, including tools for (i) visualising large multivariate data organised using hierarchical grid maps; (ii) searching for time series with patterns that can be used for making data-informed predictions; (iii) analysing contact tracing policies through simulation and visualization (Sondag et al. CGF, 2022); and (iv) analysing the parameter-sensitivity of epidemiological models with integrated algorithms and visualization (Rydow et al. TVCG, 2023, Firat et al. VI, 2022). (5) We have new methodologies for visualization research, including (i) the methodology for documenting research activities and reflections through notebooks (Dykes et al. PhilTransA, 2022), (ii) analysing and categorising a large collection of visualization imagery such as dashboards (Bach et al. TVCG, 2023); and (iii) rapid development of a visualization service with both existing service technologies and new innovative approaches. Most of these are being reported in a number of papers, including one paper published in 2021, a few papers under minor revision, and several papers that are being prepared. A summary of the RAMPVIS activities can be found at (Chen et al. Epidemics, 2022). |
| Exploitation Route | To our best knowledge, many current data infrastructures in the UK do not have adequate visualization and visual analytics capabilities. The stakeholders of these data infrastructures need to learn the lesson of COVID-19, and to make serious investment in equipping such data infrastructures with advanced visualization and visual analytics capabilities in preparation for future emergency responses. Meanwhile, the idea of developing VIS infrastructures and VIS tools for supporting data observation, data analysis, data modelling, and information dissemination has been taken forward in at least two new projects in the context of dengue outbreak prediction and monitoring and that of long-term large-scale development of machine learning models respectively. |
| Sectors | Aerospace, Defence and Marine,Agriculture, Food and Drink,Chemicals,Communities and Social Services/Policy,Construction,Creative Economy,Digital/Communication/Information Technologies (including Software),Education,Electronics,Energy,Environment,Financial Services, and Management Consultancy,Healthcare,Leisure Activities, including Sports, Recreation and Tourism,Government, Democracy and Justice,Manufacturing, including Industrial Biotechology,Culture, Heritage, Museums and Collections,Pharmaceu |
| URL | https://sites.google.com/view/rampvis/home |
| Description | (1) We have made a video "Why do we need multiple models?" which is to be distributed widely through social media to inform the public as to the importance of and challenges in epidemiological modelling. (2) We have developed a new technology for create storytelling visualization for individual regions automatically in order to provide the public with data visualization that is relevant to them. (3) We have developed a tool for allowing users around the world who do not have any modelling capability to search for COVID-19 time series similar to their regions, enabling them to observe past data with similar patterns easily and to make informed-decisions about their own regions. (4) We have developed special visualization techniques for aiding the education of school children about COVID-19 modelling in Wales. As these new developments were completed only several weeks ago, we are currently evaluating these technologies and methods while reaching out to potential users. |
| First Year Of Impact | 2022 |
| Sector | Communities and Social Services/Policy,Education,Healthcare |
| Impact Types | Societal,Policy & public services |
| Description | Dengue Advanced Readiness Tools (DART) - integrated digital system for dengue outbreak prediction and monitoring |
| Amount | £509,905 (GBP) |
| Funding ID | 226052/Z/22/Z |
| Organisation | Wellcome Trust |
| Sector | Charity/Non Profit |
| Country | United Kingdom |
| Start | 01/2023 |
| End | 12/2025 |
| Description | Long-term VIS-enabled Infrastructure for Supporting ML-assisted Human Decision-making |
| Amount | £718,638 (GBP) |
| Funding ID | EP/X029557/1 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 09/2023 |
| End | 08/2026 |
| Title | RAMPVIS Infrastructure |
| Description | The initial version of this infrastructure was developed by a team of volunteers between June and August 2020 after the team was created in response to the Royal Society's RAMP Call. The infrastructure was hosted at STFC, and it was developed in parallel with that for SCRC's data infrastructure. This infrastructure was made available online by the end of August 2020, and has been alive (at https://vis.scrc.uk/) since. Many new features have been introduced, including several tools developed during the funded period (February 2021 - January 2022). |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2021 |
| Provided To Others? | Yes |
| Impact | This R&D effort confirms that it is necessary to provide many epidemiologists and modelling science and their epidemiological modelling workflows with visual analytics support through an appropriate visualization infrastructure. Such an infrastructure should ideally be pre-existed as part of the resilience infrastructure. When such an infrastructure is not available, it is feasible to develop such an infrastructure within a relative short period. |
| URL | https://vis.scrc.uk/ |
| Description | Infrastructure and Technical Support |
| Organisation | Science and Technologies Facilities Council (STFC) |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | The RAMPVIS research activities contributed to the Royal Society's RAMP Programme (Rapid Assistance in Modelling the Pandemic), for which STFC provided critical infrastructure and technical support. |
| Collaborator Contribution | STFC provided the RAMPVIS activities (including both volunteering and funded periods) with a substantial VM Server and outstanding technical support. |
| Impact | Most research outcomes mentioned in this report were implemented on the VM server provided by STFC and benefitted from the outstanding technical support provided by STFC colleagues. |
| Start Year | 2020 |
| Title | RAMPVIS API |
| Description | RESTful APIs for the RAMPVIS system, including a data-api and an infrastructure-api. |
| Type Of Technology | Software |
| Year Produced | 2021 |
| Open Source License? | Yes |
| Impact | The data-api includes functions for data storage, and all system-level processing functions (e.g., analytical algorithms, propagation, scheduler agents, etc.). The infrastructure-api includes all functions for the core knowledge structure (i.e., ontology and the related database operations); authentication and user management; and other infrastructure related services. It also includes functions for thumbnail generation and search index services. |
| Title | RAMPVIS Analytical Support |
| Description | Software prototypes of visual analytics tools for the RAMPVIS system. |
| Type Of Technology | Software |
| Year Produced | 2021 |
| Open Source License? | Yes |
| Impact | Correlation analysis, cluster analysis, and impact analysis. |
| Title | RAMPVIS Modelling Support |
| Description | Software tools for supporting epidemiological modelling |
| Type Of Technology | Software |
| Year Produced | 2021 |
| Open Source License? | Yes |
| Impact | A VA tool for contact and tracing, and a VA tool for ensemble data analytics. |
| Title | RAMPVIS UI |
| Description | RAMPVIS user interface implemented in React |
| Type Of Technology | Software |
| Year Produced | 2021 |
| Open Source License? | Yes |
| Impact | The frontend of the main RAMPVIS server |