Online Visualization Tool Research Articles

A comprehensive Beyond-GDP database to accelerate wellbeing, inclusion, and sustainability research

“Beyond-GDP” metrics are essential for understanding societal progress. Yet despite their importance, these metrics are scattered across various databases, hindering accessibility and interdisciplinary analysis. Addressing this gap, we present the ‘WISE database’ – the first extensive collection of important Beyond-GDP metrics organized by Wellbeing, Inclusion, and Sustainability (WISE) dimensions. The WISE database consolidates data from a variety of sources, including international institutions and academic publications. It encompasses over one million data points across 244 metrics, covering 218 countries and 61 country groupings, from specific social and environmental indicators to the main Beyond-GDP indexes, and is augmented by essential metadata. The data primarily spans from 1995 to 2015, with some metrics extending back to the 19th century. To improve accessibility and data interpretation, a user-friendly online visualization tool has been developed. The WISE database aims to foster a broader view of societal progress, facilitate research on synergies and trade-offs of wellbeing, inclusion, and sustainability, and provide a foundation for interdisciplinary research.

A next-generation, histological atlas of the human brain and its application to automated brain MRI segmentation.

Magnetic resonance imaging (MRI) is the standard tool to image the human brain in vivo. In this domain, digital brain atlases are essential for subject-specific segmentation of anatomical regions of interest (ROIs) and spatial comparison of neuroanatomy from different subjects in a common coordinate frame. High-resolution, digital atlases derived from histology (e.g., Allen atlas [7], BigBrain [13], Julich [15]), are currently the state of the art and provide exquisite 3D cytoarchitectural maps, but lack probabilistic labels throughout the whole brain. Here we present NextBrain, a next-generation probabilistic atlas of human brain anatomy built from serial 3D histology and corresponding highly granular delineations of five whole brain hemispheres. We developed AI techniques to align and reconstruct ~10,000 histological sections into coherent 3D volumes with joint geometric constraints (no overlap or gaps between sections), as well as to semi-automatically trace the boundaries of 333 distinct anatomical ROIs on all these sections. Comprehensive delineation on multiple cases enabled us to build the first probabilistic histological atlas of the whole human brain. Further, we created a companion Bayesian tool for automated segmentation of the 333 ROIs in any in vivo or ex vivo brain MRI scan using the NextBrain atlas. We showcase two applications of the atlas: automated segmentation of ultra-high-resolution ex vivo MRI and volumetric analysis of Alzheimer's disease and healthy brain ageing based on ~4,000 publicly available in vivo MRI scans. We publicly release: the raw and aligned data (including an online visualisation tool); the probabilistic atlas; the segmentation tool; and ground truth delineations for a 100 μm isotropic ex vivo hemisphere (that we use for quantitative evaluation of our segmentation method in this paper). By enabling researchers worldwide to analyse brain MRI scans at a superior level of granularity without manual effort or highly specific neuroanatomical knowledge, NextBrain holds promise to increase the specificity of MRI findings and ultimately accelerate our quest to understand the human brain in health and disease.

An Online Visual Elicitation Tool to Facilitate Conversations About Supportive Care in the Context of Cognitive Impairment.

Older adults with dementia are often left out of decision-making about their care. We developed and evaluated the use of an online visualization tool to foster discussions about future care preferences. The study provides valuable insights on methods for facilitating conversations about complex topics in the context of cognitive decline.

Research trends of bone tumor treatment with 3D printing technology from 2013 to 2022: a bibliometric analysis

ObjectiveBibliometrics was employed in this study to determine the research trends in the worldwide application of 3D printing technology to treat bone tumors over the previous 10 years.MethodsPublished from 2013 to 2022, the papers related to bone tumors treated with 3D printing were located in Web of Science Core Collection (WoSCC), PubMed, and Scopus. The screened articles were included in this bibliometric study. From these papers in WoSCC, information on annual publications, journals, keywords, countries, authors, institutions, and cited references were extracted and visualized with CiteSpace (version 6.1.R6) software to investigate the state of bone tumor treatment using 3D printing as well as research hotspots. The Carrot2 online visualization tool and Vosviewer software (version 1.6.20) were employed to visualize the publications from PubMed and Scopus, respectively, in order to ascertain the most popular research topics from both databases.ResultsA total of 606, 233, and 364 publications were obtained from WoSCC, PubMed, and Scopus, respectively, between the years 2013 and 2022. In WoSCC, the peak number of publications was found in 2021, with 145 publications published. Acta Biomaterialia (11 publications) and World Neurosurgery (10 publications) were the most prolific journals, and Biomaterials was the journal cited the most (244 times). Yong Zhou was the most productive author with 14 publications, while Kwok-Chuen Wong (69 citations) and William F Enneking, (69 citations) possessed the most citations. The country with the largest quantity of publications was China (207). Among all institutions, Shanghai Jiao Tong University produced the most publications (29). Rapid prototyping was the keyword with the strongest citation burst (4.73). ‘Reconstruction’, ‘surgery’, ‘resection’, and ‘design’ caught the significant attention of researchers. 3D-printed materials, pelvic reconstruction, mandibular reconstruction, computer-assisted surgical techniques, photothermal therapy, and in vitro experiments were recognized as hot subjects and trends in current research. The most frequently occurring topics in Scopus are not significantly different from those found in WoSCC. The most prevalent research areas in PubMed encompass implant, patient-specific, bioceramic, models, and pelvic.

Development, validation and visualization of a web-based nomogram for predicting risk of new-onset diabetes after percutaneous coronary intervention

Simple and practical tools for screening high-risk new-onset diabetes after percutaneous coronary intervention (PCI) (NODAP) are urgently needed to improve post-PCI prognosis. We aimed to evaluate the risk factors for NODAP and develop an online prediction tool using conventional variables based on a multicenter database. China evidence-based Chinese medicine database consisted of 249, 987 patients from 4 hospitals in mainland China. Patients ≥ 18 years with implanted coronary stents for acute coronary syndromes and did not have diabetes before PCI were enrolled in this study. According to the occurrence of new-onset diabetes mellitus after PCI, the patients were divided into NODAP and Non-NODAP. After least absolute shrinkage and selection operator regression and logistic regression, the model features were selected and then the nomogram was developed and plotted. Model performance was evaluated by the receiver operating characteristic curve, calibration curve, Hosmer–Lemeshow test and decision curve analysis. The nomogram was also externally validated at a different hospital. Subsequently, we developed an online visualization tool and a corresponding risk stratification system to predict the risk of developing NODAP after PCI based on the model. A total of 2698 patients after PCI (1255 NODAP and 1443 non-NODAP) were included in the final analysis based on the multicenter database. Five predictors were identified after screening: fasting plasma glucose, low-density lipoprotein cholesterol, hypertension, family history of diabetes and use of diuretics. And then we developed a web-based nomogram (https://mr.cscps.com.cn/wscoringtool/index.html) incorporating the above conventional factors for predicting patients at high risk for NODAP. The nomogram showed good discrimination, calibration and clinical utility and could accurately stratify patients into different NODAP risks. We developed a simple and practical web-based nomogram based on multicenter database to screen for NODAP risk, which can assist clinicians in accurately identifying patients at high risk of NODAP and developing post-PCI management strategies to improved patient prognosis.

Real-time forecasting of COVID-19-related hospital strain in France using a non-Markovian mechanistic model.

Projects such as the European Covid-19 Forecast Hub publish forecasts on the national level for new deaths, new cases, and hospital admissions, but not direct measurements of hospital strain like critical care bed occupancy at the sub-national level, which is of particular interest to health professionals for planning purposes. We present a sub-national French framework for forecasting hospital strain based on a non-Markovian compartmental model, its associated online visualisation tool and a retrospective evaluation of the real-time forecasts it provided from January to December 2021 by comparing to three baselines derived from standard statistical forecasting methods (a naive model, auto-regression, and an ensemble of exponential smoothing and ARIMA). In terms of median absolute error for forecasting critical care unit occupancy at the two-week horizon, our model only outperformed the naive baseline for 4 out of 14 geographical units and underperformed compared to the ensemble baseline for 5 of them at the 90% confidence level (n = 38). However, for the same level at the 4 week horizon, our model was never statistically outperformed for any unit despite outperforming the baselines 10 times spanning 7 out of 14 geographical units. This implies modest forecasting utility for longer horizons which may justify the application of non-Markovian compartmental models in the context of hospital-strain surveillance for future pandemics.

An interactive visualization and data portal tool (PALTIDE) for relative sea level and palaeotidal simulations of the northwest European shelf seas since the Last Glacial Maximum

ABSTRACTRelative sea level (RSL) predictions based on glacial isostatic adjustment (GIA) simulations and palaeotidal predictions generated by hydrodynamic models using GIA‐generated palaeotopographies are available in the published literature, and datasets are available via data repositories. However, these data are often difficult to extract for specific locations or timeslices, requiring users to request datasets from corresponding authors. To overcome the intractability of these data and to enable users to interrogate datasets themselves without requiring offline requests, we have developed PALTIDE, an online visualization tool with intuitive user interface accessible at https://shiny.bangor.ac.uk/paleotidal/. The model domain for this interactive visualization tool is the northwest European continental shelf, covering the period from the Last Glacial Maximum (LGM) to the present day, and is based on previous GIA simulations by Bradley and colleagues and hydrodynamic simulations using Regional Ocean Modelling System (ROMS) published by Ward and colleagues. The tool is developed in R and utilizes a number of packages including shiny and bslib for the frontend, and arrow, raster and the tidyverse for backend data processing. The tool enables visualizations and data downloads for RSL, tidal amplitude and tide‐dependent parameters for any location within the model domain over 1000‐year timesteps from the LGM to the present.

Multiwavelength Catalog of 10,000 4XMM-DR13 Sources with Known Classifications

We present a collection of ∼10,000 X-ray sources from the 4th XMM-Newton Serendipitous Source Catalog (4XMM-DR13) with literature-verified classifications and multi-wavelength (MW) counterparts. We describe the process by which MW properties are obtained and an interactive online visualization tool we developed.

TCGAnalyzeR: An Online Pan-Cancer Tool for Integrative Visualization of Molecular and Clinical Data of Cancer Patients for Cohort and Associated Gene Discovery.

For humans, the parallel processing capability of visual recognition allows for faster comprehension of complex scenes and patterns. This is essential, especially for clinicians interpreting big data for whom the visualization tools play an even more vital role in transforming raw big data into clinical decision making by managing the inherent complexity and monitoring patterns interactively in real time. The Cancer Genome Atlas (TCGA) database's size and data variety challenge the effective utilization of this valuable resource by clinicians and biologists. We re-analyzed the five molecular data types, i.e., mutation, transcriptome profile, copy number variation, miRNA, and methylation data, of ~11,000 cancer patients with all 33 cancer types and integrated the existing TCGA patient cohorts from the literature into a free and efficient web application: TCGAnalyzeR. TCGAnalyzeR provides an integrative visualization of pre-analyzed TCGA data with several novel modules: (i) simple nucleotide variations with driver prediction; (ii) recurrent copy number alterations; (iii) differential expression in tumor versus normal, with pathway and the survival analysis; (iv) TCGA clinical data including metastasis and survival analysis; (v) external subcohorts from the literature, curatedTCGAData, and BiocOncoTK R packages; (vi) internal patient clusters determined using an iClusterPlus R package or signature-based expression analysis of five molecular data types. TCGAnalyzeR integrated the multi-omics, pan-cancer TCGA with ~120 subcohorts from the literature along with clipboard panels, thus allowing users to create their own subcohorts, compare against existing external subcohorts (MSI, Immune, PAM50, Triple Negative, IDH1, miRNA, metastasis, etc.) along with our internal patient clusters, and visualize cohort-centric or gene-centric results interactively using TCGAnalyzeR.

Global Research Hotspots in Venous Thromboembolism Anticoagulation: A Knowledge-Map Analysis from 2012 to 2021.

Venous thromboembolism (VTE) is a common cardiovascular disease that seriously threatens human lives. Anticoagulant therapy is considered to be the cornerstone of VTE treatment. An increasing number of studies has been updated in the VTE anticoagulation field. However, no bibliometric analyses have assessed these publications comprehensively. Therefore, our study aimed to analyze the global status, hotspots, and trends of anticoagulant therapy for VTE. The relevant literature on VTE anticoagulation published between 2012 and 2021 was retrieved and collected from the Web of Science Core Collection database. VOSviewer, Cooccurrence Matrix Builder, gCLUTO, and some online visualization tools were adopted for bibliometric analysis. A total of 15,152 related articles were retrieved. In recent years, the research output of VTE anticoagulation gradually increased. The United States was the most productive country. International cooperation is concentrated in North America and Europe; the most influential documents, journals, authors, and organizations were also from these two continents. Research hotspots mainly focus on clinical guidelines, VTE in special populations, non-vitamin K oral anticoagulants (NOACs), and parenteral anticoagulation. The research frontiers and trends include the assessment of NOACs and the antithrombotic management of VTE complicated with coronavirus disease 2019 (COVID-19). This bibliometric analysis provides a systematic overview of the VTE anticoagulation research, which will facilitate researchers to better understand the situation of VTE anticoagulation. Future studies should be dedicated to NOACs application and VTE-combined COVID-19 patients.

Digital Twinning for 20th Century Concrete Heritage: HBIM Cognitive Model for Torino Esposizioni Halls

In the wide scenario of heritage documentation and conservation, the multi-scale nature of digital models is able to twin the real object, as well as to store information and record investigation results, in order to detect and analyse deformation and materials deterioration, especially from a structural point of view. The contribution proposes an integrated approach for the generation of an n-D enriched model, also called a digital twin, able to support the interdisciplinary investigation process conducted on the site and following the processing of the collected data. Particularly for 20th Century concrete heritage, an integrated approach is required in order to adapt the more consolidated approaches to a new conception of the spaces, where structure and architecture are often coincident. The research plans to present the documentation process for the halls of Torino Esposizioni (Turin, Italy), built in the mid-twentieth century and designed by Pier Luigi Nervi. The HBIM paradigm is explored and expanded in order to fulfil the multi-source data requirements and adapt the consolidated reverse modelling processes based on scan-to-BIM solutions. The most relevant contributions of the research reside in the study of the chances of using and adapting the characteristics of the IFC (Industry Foundation Classes) standard to the archiving needs of the diagnostic investigations results so that the digital twin model can meet the requirements of replicability in the context of the architectural heritage and interoperability with respect to the subsequent intervention phases envisaged by the conservation plan. Another crucial innovation is a proposal of a scan-to-BIM process improved by an automated approach performed by VPL (Visual Programming Languages) contribution. Finally, an online visualisation tool enables the HBIM cognitive system to be accessible and shareable by stakeholders involved in the general conservation process.

Electrocardiogram-based deep learning improves outcome prediction following cardiac resynchronization therapy.

This study aims to identify and visualize electrocardiogram (ECG) features using an explainable deep learning-based algorithm to predict cardiac resynchronization therapy (CRT) outcome. Its performance is compared with current guideline ECG criteria and QRSAREA. A deep learning algorithm, trained on 1.1 million ECGs from 251 473 patients, was used to compress the median beat ECG, thereby summarizing most ECG features into only 21 explainable factors (FactorECG). Pre-implantation ECGs of 1306 CRT patients from three academic centres were converted into their respective FactorECG. FactorECG predicted the combined clinical endpoint of death, left ventricular assist device, or heart transplantation [c-statistic 0.69, 95% confidence interval (CI) 0.66-0.72], significantly outperforming QRSAREA and guideline ECG criteria [c-statistic 0.61 (95% CI 0.58-0.64) and 0.57 (95% CI 0.54-0.60), P < 0.001 for both]. The addition of 13 clinical variables was of limited added value for the FactorECG model when compared with QRSAREA (Δ c-statistic 0.03 vs. 0.10). FactorECG identified inferolateral T-wave inversion, smaller right precordial S- and T-wave amplitude, ventricular rate, and increased PR interval and P-wave duration to be important predictors for poor outcome. An online visualization tool was created to provide interactive visualizations (https://crt.ecgx.ai). Requiring only a standard 12-lead ECG, FactorECG held superior discriminative ability for the prediction of clinical outcome when compared with guideline criteria and QRSAREA, without requiring additional clinical variables. End-to-end automated visualization of ECG features allows for an explainable algorithm, which may facilitate rapid uptake of this personalized decision-making tool in CRT.

Rapid risk assessment tool (RRAT) to prioritize emerging and re-emerging livestock diseases for risk management.

Increasing globalization and international trade contribute to rapid expansion of animal and human diseases. Hence, preparedness is warranted to prevent outbreaks of emerging and re-emerging diseases or detect outbreaks in an early stage. We developed a rapid risk assessment tool (RRAT) to inform risk managers on the incursion risk of multiple livestock diseases, about the main sources for incursion and the change of risk over time. RRAT was built as a relational database to link data on disease outbreaks worldwide, on introduction routes and on disease-specific parameters. The tool was parameterized to assess the incursion risk of 10 livestock diseases for the Netherlands by three introduction routes: legal trade in live animals, legal trade of animal products, and animal products illegally carried by air travelers. RRAT calculates a semi-quantitative risk score for the incursion risk of each disease, the results of which allow for prioritization. Results based on the years 2016-2018 indicated that the legal introduction routes had the highest incursion risk for bovine tuberculosis, whereas the illegal route posed the highest risk for classical swine fever. The overall incursion risk via the illegal route was lower than via the legal routes. The incursion risk of African swine fever increased over the period considered, whereas the risk of equine infectious anemia decreased. The variation in the incursion risk over time illustrates the need to update the risk estimates on a regular basis. RRAT has been designed such that the risk assessment can be automatically updated when new data becomes available. For diseases with high-risk scores, model results can be analyzed in more detail to see which countries and trade flows contribute most to the risk, the results of which can be used to design risk-based surveillance. RRAT thus provides a multitude of information to evaluate the incursion risk of livestock diseases at different levels of detail. To give risk managers access to all results of RRAT, an online visualization tool was built.

Balancing Categorical Conventionality in Music

Research on the relationship between categorical unconventionality and popularity has produced mixed results. While many accounts suggest that unconventionality is penalized, much sociological theorizing indicates that success comes from a delicate balancing act between conventional and unconventional offerings. Using data on the genre self-classifications of over 2 million musicians and bands across the United States, the authors find broad support for this balancing act. Yet the shape it takes is also conditioned on local contexts, across both high-order complexes of musical genres and geographic space. The authors highlight the local metropolitan area characteristics that shift the relationship between unconventionality and popularity. They also create a typology of cities based on how their unconventional offerings are rewarded and punished. An online visualization tool enables further investigation of these relationships. The authors close by proposing an agenda for how to study local heterogeneity in the relationship between unconventionality and popularity.

DkNET Hypothesis Center: A Hub for FAIR Data, Online Resources and Hypothesis Generation

Discovering and learning how to use fast‐growing publicly available online bioinformatics and data resources can be challenging for bench scientists. The NIDDK Information Network (dkNET; https://dknet.org) is an open community resource information portal for biomedical researchers supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). dkNET helps researchers find resources relevant to their research and provides up‐to‐date information on new tools, services, and mandates to support reproducible science. To help bench scientists utilize available online bioinformatics tools, dkNET has launched a “Hypothesis Center”(https://dknet.org/hypothesis‐center) that brings together tutorials for a collection of online tools to assist researchers in exploring datasets and utilize online analytics and visualization tools. Our goal is to help researchers not only follow the step‐by‐step tutorial but also gain more understanding about these online tools, how to interpret the results, and encourage researchers to combine and/or reuse more online resources. Based on the questions and examples that a researcher has selected, the system suggests tool(s) and detailed tutorials based on specific use cases. For example, the Hypothesis Center phenotype‐genotype analytics module utilizes tools from the Signaling Pathways Project (SPP), a FAIR knowledgebase that allows researchers to ask questions across thousands of annotated transcriptomic and ChIP‐Seq datasets, and the Mouse Metabolic Phenotyping Center (MMPC), that provides information on mammalian phenotypes associated with specific gene manipulations in mouse models. The Hypothesis Center also provides information on tools and tutorials from external resources, including resources such as the Human Islet Research Network (HIRN) and the Type 1 Diabetes Knowledge Portal, that enables browsing, searching, and analysis of human genetic information linked to type 1 diabetes and related traits. To evaluate whether the approach of the Hypothesis Center can help educate researchers on using online resources more efficiently, we conducted beta testing with 15 students ranging from undergraduate to graduate. Students were new to the specific online bioinformatics tools represented and were asked to use the Hypothesis Center to learn how to use these tools. Feedback about usability and learning experience was collected to help improve the Hypothesis Center. Students found that the Hypothesis Center was helpful in learning to use a new bioinformatics tool and suggested that integrating additional tools and modifying the tutorial selections based on the knowledge level of the users could benefit trainees. To increase awareness of publicly available resources, dkNET also hosts a webinar series introducing helpful resources and promotes these resources via social media channels and YouTube. Through detailed tutorials in the Hypothesis Center and associated webinars, dkNET has created a virtual training environment that can assist researchers in finding and learning how to use online resources to support their research, breaking down barriers to the utilization of bioinformatics tools by bench researchers, and improve the utility of publicly available bioinformatics resources.

Population Genetics of Brook Trout in the Southern Appalachian Mountains

AbstractBroad‐scale patterns of genetic diversity for Brook Trout Salvelinus fontinalis remain poorly understood across their endemic range in the eastern United States. We characterized variation at 12 microsatellite loci in 22,020 Brook Trout among 836 populations from Georgia, USA, to Quebec, Canada, to the western Great Lakes region. Within‐population diversity was typically lower in the southern Appalachian Mountains relative to the mid‐Atlantic and northeastern regions. Effective population sizes in the southern Appalachians were often very small, with many estimates less than 30 individuals. The population genetics of Brook Trout in the southern Appalachians are far more complex than a conventionally held simple “northern” versus “southern” dichotomy would suggest. Contemporary population genetic variation was consistent with geographic expansion of Brook Trout from Mississippian, mid‐Atlantic, and Acadian glacial refugia as well as differentiation among drainages within these broader clades. Genetic variation was pronounced among drainages (57.4% of overall variation occurred among 10‐digit hydrologic unit code [HUC10] units or larger units) but was considerable even at fine spatial scales (13% of variation occurred among collections within HUC12 drainage units). Remarkably, 87.2% of individuals were correctly assigned to their collection of origin. While comparisons with fish from existing major hatcheries showed impacts of stocking in some populations, genetic introgression did not overwhelm the signal of broad‐scale patterns of population genetic structure. Although our results reveal deep genetic structure in Brook Trout over broad spatial extents, fine‐scale population structuring is prevalent across the southern Appalachians. Our findings highlight the distinctiveness and vulnerability of many Brook Trout populations in the southern Appalachians and have important implications for wild Brook Trout management. To facilitate application of our findings by conservation practitioners, we provide an interactive online visualization tool to allow our results to be explored at management‐relevant scales.

ZOVER: the database of zoonotic and vector-borne viruses.

Emerging infectious diseases significantly threaten global public health and socioeconomic security. The majority of emerging infectious disease outbreaks are caused by zoonotic/vector-borne viruses. Bats and rodents are the two most important reservoir hosts of many zoonotic viruses that can cross species barriers to infect humans, whereas mosquitos and ticks are well-established major vectors of many arboviral diseases. Moreover, some emerging zoonotic diseases require a vector to spread or are intrinsically vector-borne and zoonotically transmitted. In this study, we present a newly upgraded database of zoonotic and vector-borne viruses designated ZOVER (http://www.mgc.ac.cn/ZOVER). It incorporates two previously released databases, DBatVir and DRodVir, for bat- and rodent-associated viruses, respectively, and further collects up-to-date knowledge on mosquito- and tick-associated viruses to establish a comprehensive online resource for zoonotic and vector-borne viruses. Additionally, it integrates a set of online visualization tools for convenient comparative analyses to facilitate the discovery of potential patterns of virome diversity and ecological characteristics between/within different viral hosts/vectors. The ZOVER database will be a valuable resource for virologists, zoologists and epidemiologists to better understand the diversity and dynamics of zoonotic and vector-borne viruses and conduct effective surveillance to monitor potential interspecies spillover for efficient prevention and control of future emerging zoonotic diseases.

Interactive, on-line visualization tools to measure and drive equity in COVID-19 vaccine administrations.

Equitable distribution of vaccines is necessary to ensure those at highest risk of illness are protected from COVID-19 (coronavirus disease 2019). Unfortunately, there is significant evidence that vaccines have not been reaching the most vulnerable. At our large hospital system, we created interactive online tools to measure and visualize equitability of vaccine administrations and to help stakeholders identify populations at highest risk within state-designated eligible vaccine groups. Using race, ethnicity, gender, and social vulnerability, we are able to measure and reflect our vaccine administration performance against the communities that we serve. With our visualization tools, stakeholders have been able to target interventions to improve equity in vaccine administrations, including improvements in race, ethnicity, and social vulnerability. We plan to use the data elements incorporated in our electronic health record and data warehouse due to the COVID-19 pandemic to guide further population health efforts at decreasing disparities.

A holistic approach to interpretability in financial lending: Models, visualizations, and summary-explanations

Lending decisions are usually made with proprietary models that provide minimally acceptable explanations to users. In a future world without such secrecy, what decision support tools would one want to use for justified lending decisions? This question is timely, since the economy has dramatically shifted due to a pandemic, and a massive number of new loans will be necessary in the short term. We propose a framework for such decisions, including a globally interpretable machine learning model, an interactive visualization of it, and several types of summaries and explanations for any given decision. The machine learning model is a two-layer additive risk model, which resembles a two-layer neural network, but is decomposable into subscales. In this model, each node in the first (hidden) layer represents a meaningful subscale model, and all of the nonlinearities are transparent. Our online visualization tool allows exploration of this model, showing precisely how it came to its conclusion. We provide three types of explanations that are simpler than, but consistent with, the global model: case-based reasoning explanations that use neighboring past cases, a set of features that were the most important for the model's prediction, and summary-explanations that provide a customized sparse explanation for any particular lending decision made by the model. Our framework earned the FICO recognition award for the Explainable Machine Learning Challenge, which was the first public challenge in the domain of explainable machine learning.11Authors are listed alphabetically.

DPUK Cohort Explorer: An Interactive Visualisation Tool

IntroductionThe Dementias Platform UK (DPUK) Cohort Explorer is an interactive, online visualisation tool that allows users to explore data for a number of DPUK cohorts. Over 30 variables across cohorts have been harmonised, including information on demographics, lifestyle, cognition, health, and genetic biomarkers.&#x0D; Objectives and ApproachThe tool has been developed to complement existing DPUK cohort metadata to provide a visual representation of participant numbers and field-level information for a selection of cohorts. This enables users to determine a cohort’s eligibility before applying for access to a cohort’s data, and aid in shaping potential hypotheses. Developed using Microsoft PowerBI, the Explorer hosts a subset of the cohort’s baseline, harmonised data, allowing a user to interrogate the visualisations of the uploaded data in a secure manner on the DPUK Data Portal website. Visualisations are linked so that participant numbers and distributions can be explored interactively.&#x0D; ResultsThis approach allows the user to explore the harmonised data across a number of cohorts simultaneously whilst setting and adjusting filters that are of interest to the user’s search criteria. This provides a better understanding of the real-world data and enables the user to determine the feasibility of each cohort for potential studies, whilst facilitating meaningful comparisons across cohorts. The tool currently visualises five DPUK cohorts with a total of 82,391 participants, however it is being incrementally developed with more cohorts being added continually.&#x0D; Conclusion / ImplicationsBy combing an easy-to-use, interactive dashboard with harmonised sets of real-world data, the tool allows the user to explore, interrogate and better understand field-level information in a secure manner with zero data transfer. This provides more insight for the user when applying for access to a cohort dataset using the DPUK Data Portal and may help the user to make more informed decisions and/or hypotheses.