The research, led by Victor de Brito of the Cornell University Museum of Vertebrates, addresses a persistent problem in the museum world. Collections serve research, education, conservation, and public outreach, and a central goal is making their growing volumes of data accessible to researchers, educators, and the public. In most cases that access happens through institutional websites and web applications, but limited access to programming expertise, dedicated personnel, and large-scale data management infrastructure remains a major challenge for many collections.

Large-scale aggregators such as the Global Biodiversity Information Facility, VertNet, and MorphoSource have substantially improved data availability across institutions. However, the study notes that these platforms often lack the most up-to-date information, because collections upload their data in episodic batches. They are also vulnerable to shifts in funding policies that may limit hosted data or impose new fees on partner institutions. Aggregators are generally not tailored to the specific needs of individual collections, and the hosting institutions are often overlooked in acknowledgments by authors of downstream analyses.

Recent advances in generative artificial intelligence, particularly large language models that understand and generate human-like text, have transformed how developers create, maintain, optimize, and test code. The scale of adoption is striking: GitHub, the largest repository for software development projects, added more than 36 million developers in 2025 and reached over 180 million users, a surge largely attributed to AI coding tools. A survey found that 90 percent of software developers at Google reported using AI in their work, and 71 percent of those who write code used AI-assisted coding. By translating natural language instructions into executable code, these systems substantially lower the technical barrier to software creation.

To test whether that barrier has fallen far enough for museum professionals, the researcher developed two applications using GPT-5.3-Codex on Windows 11 through Visual Studio Code, with a medium reasoning effort setting. Both were built from .csv files exported from the museum’s Specify 7 database, tested locally, and then deployed online using GitHub Pages. The repositories included an AI-generated code disclosure statement and an MIT License file clarifying usage and distribution permissions.

The first application is a specimen search portal for the museum’s ichthyology collection. Built with React, TypeScript, Vite, Tailwind CSS, Leaflet for mapping, and PapaParse for CSV parsing, it allows users to filter specimen records by order, family, genus, species, catalog number, preparation type, date, and geography. Filters support dropdown lists with autocomplete, multi-select, and dynamic interdependence, meaning that selecting one value updates all other dropdowns to show only compatible options. Users can export filtered results as a new .csv file and view an interactive map of sampling localities that updates live as filters change, with clicking a map point highlighting the corresponding table row.

The second application is an interactive visualization of sampling localities across the museum’s four vertebrate collections, covering mammals, birds, reptiles and amphibians, and fishes. Color-coded, semi-transparent points distinguish the collections so overlap density remains visible, and an animated timeline makes points appear year by year from earliest to latest. Playback controls include play and pause, speed settings, year navigation, a loop toggle, and a draggable year slider. Below the map, a Plotly line chart displays cumulative specimen and lot counts over time, with one line per collection and a black total line, updating dynamically as the selected year changes.

Validation showed the tools perform accurately. The specimen search portal produced results consistent with the museum’s existing Specify Web Portal across all 50 randomly selected genus and species combinations tested, with query results exported as .csv files and compared in R to identify any catalog numbers present in one portal but absent in the other. The interactive map accurately reproduced specimen localities when visually compared with coordinates exported from the Specify database and plotted in QGIS. Once the data structure was exported, functional applications were generated and tested in less than 30 minutes, allowing staff to rapidly test different ways of presenting and exploring their data.

The study also weighs alternatives and caveats. Low-code and no-code platforms allow tool development with minimal programming, but they constrain customization, lock users into platform-specific architectures, and carry subscription costs. AI-assisted coding instead generates fully editable source code, offering greater adaptability as taxonomies, digitization priorities, and data standards evolve. However, large language models are stochastic by design, sampling from probability distributions, so results varied with each execution even when the same prompt was repeated. Clearly structured prompts describing interface elements, functionality, data structure, and outputs in detail greatly improve code quality, and multi-step refinement remains essential for producing fully functioning applications.

Several practical considerations accompany adoption. AI-assisted coding works best when users have some programming knowledge, since they remain responsible for troubleshooting, maintenance, and deployment, effectively acting as their own technical support. Complexity raises the skill requirement: both applications here operate from exported .csv files and are tightly coupled to the original data structure, so connecting to a live server or database would require substantial refactoring. Data security is critical, and sensitive information should never appear in publicly accessible applications or connect directly to live institutional databases; exported datasets or copies reduce breach risks. Access to AI platforms may require paid subscriptions, though applications can be developed quickly and maintained independently afterward, minimizing long-term expense. The broader challenge now shifts from implementation to identifying meaningful tools and sharing them in standardized ways across institutions, since rapidly generated customized tools risk fragmenting database ecosystems. With careful prompt design and attention to interoperability standards such as Darwin Core, the approach could transform collection staff from end-users of pre-built systems into active designers of their own digital infrastructure.

Subject of Research: AI-assisted coding for developing digital tools in natural history collections

Article Title: Using AI-assisted coding to build digital tools for natural history collections

Article References: Using AI-assisted coding to build digital tools for natural history collections. (n.d.). https://doi.org/10.3897/nhcm.3.191202

Image Credits: AI Generated

DOI: 10.3897/nhcm.3.191202

Keywords: AI-assisted coding, natural history collections, large language models, GPT-5.3-Codex, Cornell University Museum of Vertebrates, specimen data, digitization, web applications, data visualization, GBIF, Specify, museum informatics

The project, called Who Eats Whom, was launched in 2019 on iNaturalist, one of the world’s largest citizen science platforms, which holds hundreds of millions of geolocated, time-stamped photographs of wildlife. The idea is elegantly simple. Volunteers who happen to photograph an animal in the act of feeding, whether a glaucous-winged gull tearing into a starfish, a giraffe browsing a duiker-berry tree, or an American crocodile snatching an invasive lionfish, can submit their images to the project. Each feeding event generates two linked observations, one for the consumer and one for the consumed, and the iNaturalist community collectively proposes and refines the taxonomic identifications for both. Metadata tags then bind the pair together as a documented trophic interaction, complete with photographic evidence.

The scale of participation has grown far beyond what the founders initially promoted. Since its launch, Who Eats Whom has amassed nearly 14,000 observations of feeding interactions contributed by close to 2,000 observers spread across more than 100 countries. Remarkably, most of these contributions have come from volunteers who discovered the project on their own while browsing the iNaturalist database, rather than through targeted recruitment campaigns. This organic growth suggests that the photographic bycatch of ordinary biodiversity observation, images of feeding that people capture incidentally while documenting species, represents a vast and largely untapped reservoir of ecological information covering a geographic and taxonomic breadth that no targeted field study could ever achieve.

In 2025, the team began building a web application to transform this raw trove of observations into a usable scientific resource, releasing it in early 2026. The platform draws on the iNaturalist Application Programming Interface to collate data in real time, meaning the database grows dynamically as new observations pour in. The site restricts itself to so-called research-grade observations, those in which the iNaturalist community has reached high agreement on the identity of both species involved. The developers are careful to note that research-grade status confirms the taxonomy but does not independently verify that feeding actually occurred in every image, a distinction that matters for rigorous use. Even so, the filtering yields a network that, as of January 2026, connects 1,863 species through 1,607 unique feeding interactions, a figure the team expects to climb toward hundreds of thousands as iNaturalist itself continues its rapid expansion.

The technical architecture of the platform is designed around accessibility and scalability. Users type in a species of interest and can ask either what it eats or what eats it, then explore the results through five complementary visualization modes. A grid view displays the underlying photographs, putting faces, so to speak, on every interaction. A graph view and a network view render the data as directed networks in which nodes represent species and arrowheads point toward the consumer, with line widths encoding how frequently an interaction has been documented. A map view plots interactions geographically, while a global view renders the entire database as one sprawling interactive food web. Every search can also be exported as a CSV file, lowering the barrier for researchers who want to fold the data into their own analyses.

For ecologists, the applications are numerous. The database can supply basic natural history information about the dietary needs or natural enemies of poorly studied species, a gap that remains enormous given that the feeding ecology of most organisms on Earth has never been formally documented. It can surface rare or previously unrecorded interactions, flag highly connected prey species that may serve as keystone food sources, and expose taxa or regions where data are so sparse that targeted fieldwork is warranted. Conservation biologists could use it to model the potential trophic ripple effects of an invasive species or to search for candidate biological control agents. The temporal and spatial metadata attached to each observation open further avenues, from tracking how diets shift across seasons and life stages, a field known as feeding phenology and ontogeny, to cataloging which interactions currently exist so that they can be protected before they vanish.

The researchers are candid about the limitations inherent in unstructured citizen science data. Because observations are not collected under controlled sampling protocols, the database inevitably mirrors the biases of the people behind the cameras. Large, conspicuous, charismatic, and terrestrial organisms are overrepresented, while small, cryptic, aquatic, and nocturnal interactions remain comparatively rare. Who Eats Whom is therefore best understood as a living natural history archive rather than a representative sample of the planet’s trophic relationships. Even in well-studied and relatively simple ecosystems, food webs are extraordinarily complex, and constructing complete ones has proven stubbornly difficult. Yet incomplete food webs still carry genuine scientific value, particularly when analyzed with quantitative network methods, and centralized repositories like this one create the capacity to merge photographic evidence with complementary techniques such as DNA metabarcoding, producing more robust and multi-layered pictures of ecological networks than any single method could deliver alone.

Beyond the laboratory, the project has been deliberately shaped as a tool for education and public engagement. Food webs are a staple of secondary school science curricula, and questions about what animals eat resonate far beyond academia, whether a gardener wondering which plants best support native pollinators or a hiker curious about the predator lurking at the top of a local trail. The team is co-developing guidelines with the iNaturalist community for classifying the different types of feeding interactions, distinguishing parasitism from herbivory, nectar feeding from scavenging, and plans future updates including improved search queries, richer network visualizations, a dedicated R package for programmatic access, and a community vetting system for data quality. With an organized corps of volunteers reviewing submissions in real time and the sophisticated iNaturalist identification machinery working underneath, the database is effectively self-improving.

There is a philosophical dimension to the project as well. When John Muir wrote that trying to pick out anything by itself reveals it hitched to everything else in the universe, he was describing precisely the relational view of nature that Who Eats Whom seeks to operationalize. The wonder inspired by organisms such as Maculinea caterpillars, which mimic the pheromones and even the sounds of ant queens to be adopted into colonies and fed by worker ants, lies not in their existence as isolated species but in the astonishing specificity of their relationships. By shifting attention from species in isolation to the diversity of connections among them, the project offers a corrective to a reductionist habit in Western science and makes those connections visible and tangible to anyone with a web browser.

What began as a modest effort to harvest the accidental bycatch of wildlife photography has matured into the first global food web built from verifiable citizen science images, a complement to existing interaction databases such as Global Biotic Interactions and a citizen science project in its own right. Its creators anticipate that letting the public visualize and manipulate the tangled web of who consumes whom could deepen appreciation for the interconnected natural world and, perhaps, strengthen the resolve to conserve it. In the meantime, every hiker, gardener, and beachcomber with a smartphone holds the potential to add another thread to the map, one photograph of lunch in the wild at a time.

Subject of Research: A citizen science-derived global database of feeding interactions between species built from iNaturalist photographs

Article Title: Who Eats Whom? A global food web derived from citizen science

Article References: Allf, B. C., Mallavarapu, A., Kikuchi, D. W., Vasudeva, N., & Dunn, R. R. (2026). Who Eats Whom? A global food web derived from citizen science. PLOS Biology, 24(9), e3003988. https://doi.org/10.1371/journal.pbio.3003988

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3003988

Keywords: citizen science, food webs, iNaturalist, species interactions, trophic ecology, biodiversity, ecological networks, predation, natural history, conservation, data visualization, public engagement

The motivation for the tool stems from a persistent gap in how statistical software presents regression. According to the authors, mainstream packages such as SPSS, Minitab, Tableau, and JMP present regression results through disconnected numerical tables and static plots, obscuring the geometric relationships that are essential for deep conceptual understanding. A systematic survey of ten widely used systems conducted by the developers found that none of them provides geometric error decomposition or sampling-based interval exploration, while coordinated diagnostic views appear only in limited or partial form in a small number of systems. RegVis was built specifically to close those three gaps within a single, installation-free web application.

The first and perhaps most striking feature is the geometric decomposition of variance. For every data point, RegVis draws a red translucent square whose side length equals the absolute residual, the distance between the observed value and the fitted value predicted by the regression line. It also draws a blue translucent square whose side length equals the distance between the observed value and the mean of the response. The aggregate area of all red squares corresponds to the sum of squared errors, SSE, while the aggregate area of all blue squares corresponds to the total sum of squares, TSS. The identity R² = 1 − SSE/TSS, usually presented as an abstract algebraic formula, thus becomes a visible contrast between two fields of colored area. When a user drags a point or adds an outlier, the squares animate smoothly, and the shift in the ratio of red to blue area makes the change in explained variance perceptually apparent rather than merely calculable.

The developers are careful to note that this area-based rendering is intended as a conceptual aid rather than a precise measurement instrument. Research in graphical perception going back to the classic work of Cleveland and McGill shows that area encodings are less accurate than position or length for exact quantitative comparison. RegVis embraces this limitation deliberately: the squares are meant to convey a holistic, intuitive sense of how total variance splits into explained and unexplained components, while the exact numeric value of R² remains available in a live readout at all times. Users can toggle each layer independently and limit the maximum number of displayed squares to reduce visual clutter, supporting a progressive disclosure approach that starts with a simple scatterplot and builds toward the full decomposition.

The second major innovation tackles one of the most misunderstood concepts in statistics: the confidence interval. Studies have shown that even professional researchers harbor robust misinterpretations of what a confidence interval actually means. RegVis addresses this with an interactive sampling demonstration. Clicking the Sample button draws a random subsample of a user-configured proportion from the current dataset, fits a least-squares line to that subset, and overlays the result as a persistent semi-transparent grey trace. Each click produces a fresh, independent sample, and the accumulating cloud of grey lines gives learners a visceral sense of sampling variability. The displayed confidence band, computed once from the complete dataset, serves as the reference against which the repeated sample lines can be judged.

Because the dataset is treated as a finite population, the behavior of the containment rate follows finite-population sampling theory rather than the textbook infinite-population ideal. A grey trace counts as contained only if it lies within the confidence band at every value of x. As the sampling fraction increases, sampling variance shrinks through the finite-population correction, and containment rises monotonically toward certainty at 100 percent, where the subsample coincides with the population. To make this relationship observable rather than merely asserted, a Simulate function draws 100 independent samples at each of eleven sampling fractions from 0 to 100 percent, fits a line to each, and plots the resulting containment rate against the sampling fraction. In one illustrative workflow described in the paper, a 50-point dataset at the 95 percent confidence level produced a curve rising from roughly 10 percent containment at a 10 percent sampling fraction to full containment at 100 percent.

Underlying both features is a coordinated three-view architecture designed around Shneiderman’s direct-manipulation principle. The primary scatterplot displays the data, the red regression line, the toggleable error squares, adjustable confidence and prediction intervals, and a live regression equation with R² and sample size. A residuals-versus-fitted-values panel reveals heteroscedasticity as cone-shaped patterns, non-linearity as systematic curves, and outliers as extreme vertical displacements. A Q-Q plot maps empirical residual quantiles against theoretical normal quantiles, exposing heavy tails, skewness, and individual aberrant points. Hovering over any observation in any view triggers synchronized highlighting across all three panels, allowing users to trace a single data point through data space, error space, and distributional space simultaneously. Clicking anywhere adds a point; clicking an existing point removes it; and every such action triggers the full pipeline of data update, statistical recalculation, and rendering across all three views.

That pipeline is fast. The developers instrumented each architectural layer with performance timers and ran 30 repetitions at six dataset sizes across Microsoft Edge, Firefox, and Chrome on a 12-core Windows desktop. The computation layer, which recalculates least-squares coefficients, residuals, SSE, TSS, R², and interval bounds from scratch after every change, stayed below 1.2 milliseconds on average across all tested conditions. End-to-end latency rose gradually from under 2.3 milliseconds at ten points to a maximum of 17.93 milliseconds at 1,000 points in Firefox, comfortably below the roughly 100-millisecond threshold at which interactions are typically perceived as instantaneous. Rendering dominates the total, contributing an order of magnitude more time than computation even at the largest tested size, and the authors note that performance was not evaluated beyond 1,000 points. A benchmark script ships with the repository so that others can verify or extend the measurements.

The system is implemented entirely in client-side JavaScript using D3.js version 7, with HTML5 and CSS3, and requires no server-side installation; it can be deployed as a static file on any web host or opened directly from a local filesystem. The choice of D3.js over higher-level charting libraries was deliberate, since the custom geometric error squares, the sampling accumulation overlay, and the synchronized cross-view highlighting all demand low-level control over SVG elements that pre-built statistical libraries do not expose. The software is released under the MIT license on GitHub, includes a guided five-step tutorial covering dataset construction, variance decomposition, interval sampling, and diagnostic coordination, and supports CSV import and export with basic validation for reproducibility.

An exploratory usability assessment with eighteen expert participants, each holding at least one graduate-level statistics course, yielded a mean System Usability Scale score of 79.2 with a standard deviation of 12.8, a figure the authors describe as indicating high perceived usability, though they caution that the study lacked a controlled baseline and formal hypothesis testing. The developers also disclose that Ahmad is the founder of SpimeLab Inc., which hosts the platform, while stating that the reported results were developed independently of commercial considerations. Looking forward, the design principles of geometric correspondence, observable sampling, and coordinated multi-view interaction are described as general enough to extend to multiple regression, generalized linear models, ANOVA, and cross-validation. For now, RegVis positions itself as the first comprehensive visual environment in which regression mathematics is not merely calculable but directly observable, one clickable point at a time.

Subject of Research: Interactive web-based visualization of linear regression analysis, variance decomposition, and confidence interval sampling for statistics education

Article Title: RegVis: An interactive web-based visualization system for linear regression analysis and diagnostics

Article References: Ahmad, U., & Uzair, H. (2026). RegVis: An interactive web-based visualization system for linear regression analysis and diagnostics. SoftwareX, 36, Article 103045. https://doi.org/10.1016/j.softx.2026.103045

Image Credits: AI Generated

DOI: 10.1016/j.softx.2026.103045

Keywords: linear regression, data visualization, statistics education, confidence intervals, D3.js, variance decomposition, open-source software, residual diagnostics, interactive learning, R-squared, sampling variability, web application

The new paper is the third and last part of what the author describes as a research triptych. The first two installments dealt with box score data and play-by-play data respectively, each feeding into a growing analytical infrastructure for Spanish basketball. Box scores capture cumulative statistics such as points, rebounds, and assists, while play-by-play records describe the sequence of events during a game. Shooting data, however, adds a crucial spatial dimension. For each attempt, the dataset records the location of the shot on the court, the shot type, and the outcome. This is precisely the kind of information that has transformed how analysts, coaches, and fans understand the modern game, from the rise of the three-point shot to the strategic value of shots near the rim.

To build the dataset, Vinué developed code in the R programming language that automatically collects shooting information from the official ACB website using a web scraping procedure. Web scraping involves programmatically extracting data from web pages, and the author designed the procedure to be friendly and reproducible. The approach draws on established R tools for harvesting web content, including packages such as rvest and RSelenium, and respects the target website’s robots.txt file, which specifies rules for automated access. Once collected, the raw data required ordinary cleaning to be structured into a standard format suitable for analysis. The author then validated the accuracy of the scraped data by plotting the shots and comparing their coordinates and outcomes against the shot charts displayed on the source website, ensuring that the automated pipeline faithfully reproduces the original information.

The scale of the resulting dataset is substantial. It covers the 2024-2025 ACB season and includes 39,550 shots drawn from 306 games and 293 players. Each record contains, among other details, the location of the attempted shot, the shot type, and whether the attempt was successful. A dataset of this size allows for detailed shooting profiles of individual players, showing where on the floor they prefer to shoot and how effective they are from each area. Such profiles go far beyond a simple field goal percentage, which collapses all attempts into a single number regardless of where they were taken. Spatially aware statistics have become central to basketball analytics worldwide, and this work brings that capability to Spanish league data in a systematic, reproducible way.

A key contribution of the study is the integration of shooting data into a previously developed data platform, so that all analyses are centralized in a single place. The platform, described in an earlier paper, already handled box score and play-by-play information; adding shooting data completes the triptych and turns the platform into a comprehensive resource for Spanish professional basketball. Alongside the platform, the author updated BAwiR, an accompanying R package that is freely available from CRAN, the comprehensive archive network for R software. The package includes documentation for all its files as well as three vignettes that guide users through its functionality. Making both the data and the code openly available means that other researchers, analysts, and enthusiasts can reproduce the results and extend the analysis to their own questions.

To make the data accessible to non-programmers, the study also provides a web application for easy interaction. Built with Shiny, a popular framework for turning R code into interactive web tools, the application lets users explore shooting data without writing any code themselves. Interactive visualization has become an essential bridge between raw data and decision-makers in sport, allowing coaches and analysts to query specific players, games, or shot types on demand. The author also gathered feedback from users through a questionnaire, reflecting a growing recognition in the visualization research community that sports tools should be designed in consultation with the people who will actually use them.

One of the more technically interesting parts of the work concerns how to divide the basketball court into meaningful zones. Classical zones such as the paint, the mid-range areas, and the corners are usually defined manually, but the paper explores whether automatic clustering algorithms can do the job instead. Three methods were tested on the shooting dataset: DBSCAN, HDBSCAN, and k-means. DBSCAN, which stands for Density-Based Spatial Clustering of Applications with Noise, groups points that lie close together in space and marks isolated points as noise, but it requires the user to specify a neighborhood radius and a minimum number of points in advance. HDBSCAN extends this idea by building a hierarchy of clusters of varying density, reducing the need to tune the radius parameter. K-means, the simplest of the three, iteratively assigns each shot to its nearest cluster center and then recalculates those centers as the average of the assigned points, with the number of clusters chosen beforehand, often through silhouette analysis.

The results of the clustering comparison were instructive. K-means produced the closest representation of the classical basketball regions, identifying six zones, although it did not distinguish between two-point and three-point shots, and some regions contained both. HDBSCAN roughly identified some areas, such as the paint and the left and right three-point zones, but its other regions were messy. DBSCAN fared worst, returning just two large regions that carry little basketball meaning. When the analysis was refined by filtering first by shot type, k-means divided the two-point area into three regions approximating the paint and the left and right mid-ranges, while the three-point area split into six clearly identifiable regions: left and right corners, left and right elbows, center three-point shots, and long-distance three-point attempts. These regions largely coincide with the manual partition proposed by the author, and the clustering even suggested numerical limits for the court zones, such as paint boundaries at y-coordinates of around plus or minus 2000. The conclusion is that k-means can serve as a useful starting point for determining the coordinates needed to build court regions manually.

The study is candid about the limitations of automated data collection. The scraping code is reproducible with similar shot data, but if the target website changes its structure, manual updates and further cleaning may be required. This is a familiar trade-off in sports analytics, where researchers often depend on websites they do not control. The author acknowledges the official ACB website for making the data accessible, and the paper’s appendices provide a step-by-step guide to the data collection strategy, along with a glossary of basketball terms to help readers less familiar with the sport’s vocabulary. Such transparency supports reproducibility, a value increasingly emphasized in the R sports analytics community, where packages for basketball, football, and other sports are systematically reviewed and shared.

The broader significance of the work lies in democratizing access to professional basketball data outside the NBA. While the NBA offers well-documented application programming interfaces that have spawned a rich ecosystem of analytical tools, European leagues have historically lagged in data availability. Research has shown that analytics investment can influence team performance in professional basketball, and systematic reviews have highlighted both the potential and the uneven adoption of analytical techniques in the sport. By centralizing box score, play-by-play, and shooting data for the ACB in one platform, backed by a free R package and an interactive web application, this research gives Spanish basketball a data infrastructure comparable in spirit to what NBA analysts enjoy. From shot charts and spatial clustering to player shooting profiles, the tools are now in place for coaches, journalists, and fans to interrogate the game at a level of detail that was previously out of reach for Spain’s top league.

Subject of Research: Web scraping and visualization of shooting data from the Spanish ACB basketball league

Article Title: Adding shooting data to a centralized platform for basketball data visualization

Article References: Adding shooting data to a centralized platform for basketball data visualization. (n.d.). https://doi.org/10.1007/s11042-026-21930-2

Image Credits: AI Generated

DOI: 10.1007/s11042-026-21930-2

Keywords: basketball analytics, ACB, shooting data, web scraping, R package, data visualization, Shiny web application, k-means clustering, DBSCAN, HDBSCAN, shot charts, sport analytics

The work, carried out by Neelu Verma and Anand Mishra of the Indian Institute of Technology Jodhpur, addresses a gap that has persisted in automated fact-checking research. Existing benchmarks for chart-based verification typically assume the relevant chart is already provided to the system, a convenience that rarely matches reality. In genuine settings—a journalist checking a viral post, an analyst auditing a report—a claim may refer to one chart among thousands circulating in a document archive or across the web. The researchers frame the task accordingly: their framework must first retrieve the correct chart from a large collection and only then decide whether the claim is supported or refuted by what that chart depicts.

To make such a setting trainable and measurable, the team built FactViz3M, a dataset whose scale dwarfs prior efforts. It contains 364,000 chart images and 29.5 million factual assertions, making it 23 times larger than existing datasets in images and a striking 1,867 times larger in factual statements. The dataset was assembled from publicly available sources, with claims generated in a controlled way. Supported facts were produced by transforming chart question-and-answer pairs into declarative statements, while each supported fact was paired with a refuted counterpart generated from the chart image itself, guided by carefully designed examples.

The construction of the refuted claims is technically noteworthy because it avoids the trap of obviously false statements. The generation process perturbs numerical values by roughly 5 to 20 percent relative to the original, preserving units, formatting, and plausible magnitude so that the false claim could reasonably appear on a chart. For comparative claims, the perturbations flip relationships rather than invent content: greater-than becomes less-than, highest becomes lowest, increased becomes decreased. The result is a benchmark of challenging, realistic misinformation in which the falsehood lies in subtle details rather than in implausible fabrications.

On the architectural side, the proposed framework operates in stages that mirror how a careful human fact-checker might work. Given a textual claim and a pool of charts, the system first retrieves candidate charts relevant to the claim. It then decomposes the claim into atomic sub-facts—elementary assertions such as a single value, a single comparison, or a single trend—rather than trying to evaluate the entire statement at once. Each sub-fact is verified against the retrieved chart using a vision–language model, a class of neural networks trained jointly on images and text that can read visual content and reason about it linguistically.

The final stage aggregates these piecewise judgments through self-consistent reasoning, a technique in which the model’s conclusions are checked across multiple reasoning passes so that only stable, reproducible verdicts survive. This decomposition-and-aggregation strategy serves two purposes. It improves accuracy, because verifying small atomic facts is more reliable than asking a model to judge a compound claim in one shot, and it yields interpretability: the intermediate reasoning steps expose exactly which sub-facts were supported and which were refuted, allowing a human user to audit why the system reached its conclusion.

The experimental results show the approach outperforming strong baselines, achieving an F1 score of 76.7 percent, roughly 2 percentage points better in absolute terms than competing methods. The F1 score, which balances precision and recall into a single measure, is the standard yardstick for verification tasks where both false positives and false negatives carry costs. A 76.7 percent F1 on a task combining retrieval and fine-grained visual reasoning represents meaningful progress, though the number also makes clear that chart-grounded verification at scale remains an open challenge rather than a solved problem.

Why does this matter beyond the benchmark? Misleading charts are a staple of online misinformation precisely because they are hard to check quickly. A reader confronting a viral graphic showing, say, a supposed collapse in some economic indicator has no easy way to know whether the underlying data support the claim, or whether the chart even corresponds to a real dataset. Systems of the kind described in this paper point toward automated tooling that could trace a claim back to its source chart, decompose the assertion, and report which parts hold up. The researchers’ emphasis on interpretable intermediate steps is particularly relevant here, since fact-checking outputs that cannot be explained are unlikely to earn public trust.

The study also situates itself within a broader movement in artificial intelligence known as retrieval-augmented reasoning. Rather than expecting a model to hold all relevant knowledge in its parameters, retrieval-augmented systems fetch external evidence—documents, tables, or in this case chart images—and ground their reasoning in that evidence. Verma and Mishra had previously explored related territory with work on natural-language query-to-chart image retrieval, and the new paper extends that line by coupling retrieval with verification. The approach draws on recent advances in vision–language models such as Qwen2-VL and related multimodal systems, which have dramatically improved machines’ ability to read axes, legends, and data points embedded in visualizations.

Limitations and open questions remain. The dataset is curated from publicly available sources, and claims are generated synthetically, which means the system’s performance on organically produced misinformation—crafted by humans with intent to deceive in messier, more creative ways—still needs real-world testing. The 76.7 percent F1 score leaves nearly a quarter of the verification signal imperfect, and retrieval errors can propagate: if the wrong chart is fetched, even perfect reasoning about it cannot save the verdict. Still, by defining the retrieval-augmented version of chart fact-checking and supplying a dataset three orders of magnitude larger than its predecessors, the IIT Jodhpur team has given the field a concrete foundation. As charts continue to multiply across the information ecosystem, tools that can ask, find the right picture, and check the numbers against it may become as essential to the fact-checking pipeline as text-based verification has been for the past decade.

Subject of Research: Retrieval-augmented fact verification of claims against chart images using vision–language models

Article Title: Retrieval-augmented chart-grounded fact verification with self-consistent vision–language reasoning

Article References: Verma, N., & Mishra, A. (2026). Retrieval-augmented chart-grounded fact verification with self-consistent vision–language reasoning. International Journal of Data Science and Analytics, 22(1), Article 300. https://doi.org/10.1007/s41060-026-01277-5

Image Credits: AI Generated

DOI: 10.1007/s41060-026-01277-5

Keywords: fact verification, charts, vision–language models, retrieval-augmented reasoning, FactViz3M, misinformation, dataset, self-consistent reasoning, machine learning, data visualization, artificial intelligence, chart retrieval

The project taps into iNaturalist, one of the world’s largest biodiversity platforms, where members of the public upload photos, video and audio recordings of the organisms they encounter. While iNaturalist has long been used to map where species live, the new initiative extends its power to a different layer of ecological information: what those species are actually doing when they encounter one another. Users of the platform can now explicitly tag both predator and prey within a single photograph or video, capturing a feeding interaction at the moment it happens. The predator in question may be a lion taking down an antelope or a caterpillar munching on a garden leaf; the database treats both interactions as equally valuable data points in the planetary food web.

Bradley Allf, corresponding author of the paper describing the project and currently a postdoctoral researcher at Colorado State University, began the work while a graduate student at North Carolina State University. He explains that knowledge of what organisms eat remains fundamentally incomplete and fragmented across the academic literature. iNaturalist, he notes, has already proven its worth in documenting which organisms live in different locations, and this project extends that success to feeding relationships. The result is a publicly available website, who eats whom dot org, where anyone can explore recorded trophic interactions through an easily searchable interface. What began as an effort to consolidate scattered knowledge has quickly become one of the largest structured datasets of its kind.

The numbers are already striking. To date, the Who Eats Whom project has logged approximately seventeen thousand observations of trophic interactions involving roughly five thousand seven hundred species. These records have come from around two thousand observers based in more than one hundred countries, spanning ecosystems from tropical forests to suburban backyards. Each observation represents a moment when a person witnessed and documented one organism consuming another, creating a verifiable record that researchers can analyze, visualize and cross-reference with existing scientific literature. In aggregate, these individual moments begin to sketch the outline of a global food web, a planetary-scale network of feeding dependencies that has never been mapped at this level of public participation.

Aditi Mallavarapu, co-author of the paper and an assistant professor of computer science at NC State, approached the project from the perspective of human-computer interaction and tool design. Her interest, she says, lies in building usable interactive tools and learning environments, and the food web presented an irresistibly complex system to work with. The design goals were explicit: the site had to be usable, visually engaging and ethically responsible in the way it handled data contributed by citizen scientists. Mallavarapu emphasizes that the platform is intended to serve both researchers and the general public, helping each group grasp feeding relationships that are too numerous and too intricate to be understood through static tables or technical papers alone.

One of the most revealing early findings concerns what motivates participants. Mallavarapu observes that public interest extends well beyond charismatic predators such as lions and wolves. Contributors are documenting smaller, quieter trophic relationships, such as the insects that feed on persimmon leaves, and are using the site to explore interactions tied to their personal passions and local environments. This breadth of engagement matters scientifically. Ecological research has historically skewed toward large, photogenic species, leaving the vast majority of feeding interactions — particularly those involving invertebrates, plants and microbes — poorly documented. A platform driven by thousands of curious amateurs naturally fills some of those gaps, recording interactions that professional ecologists would rarely have the time or funding to chase.

From a technical standpoint, the database opens several promising research directions. Allf notes that the team can use Who Eats Whom to identify novel trophic relationships that have not yet appeared in the academic literature, then contribute those findings formally to scientific publications. The long-term ambition is for the platform to function as a parallel database that complements the peer-reviewed literature, compiling as much feeding information as possible into a single, accessible location. Because observations are time-stamped and geolocated through iNaturalist, the dataset also carries temporal and spatial dimensions that static literature reviews cannot match, potentially allowing scientists to track how feeding relationships shift across seasons, habitats and years.

Visualization is central to that ambition, and it is where Mallavarapu’s team is investing heavily. She describes a planned visual representation of this planetary-scale food web, designed to capture how trophic interactions relate to one another and how those relationships change over time. A simple version of the interface is already live, but the developers are building richer features that will let users explore feeding patterns and pose practical, personal questions — such as what to plant in a North Carolina garden to attract hummingbirds. That framing reveals the project’s dual identity: it is simultaneously a serious scientific database and an invitation for ordinary people to see themselves as participants in the ecological networks surrounding them.

Data quality remains a central challenge for any citizen science effort, and the team is developing tools specifically to validate submissions tagged on iNaturalist. Verification workflows will help ensure that predator and prey identifications are accurate and that interactions are recorded consistently, protecting the database’s scientific value as it scales. Beyond research, the project has an educational mission. The team has already created educational games built on the database, aimed at helping children understand the food web and the interdependencies it captures, with plans to expand that work if additional funding can be secured. The project was developed with substantial contributions from NC State students, including Maithili Bhoop, Adam Biscoe, Surabhi Nair and Lavanya Middha, alongside the thousands of iNaturalist users whose field observations make the entire enterprise possible.

The underlying research, titled Who Eats Whom? A global food web derived from citizen science, was published on September 22 in the open-access journal PLOS Biology. Senior author is Rob Dunn, a professor in NC State’s College of Agriculture and Life Sciences; additional authors include Nikhil Vasudeva, a former NC State undergraduate, and David Kikuchi, an assistant professor of integrative biology at Oregon State University. The researchers report no conflicts of interest. As the database grows, its creators hope it will transform not only how scientists study food webs but also how the public understands them — turning every hiker, gardener and backyard naturalist into a node in a worldwide effort to document the fundamental question of ecology: what eats what, and why it matters for the living planet.

Subject of Research: A citizen science platform that compiles global predator-prey interaction data from iNaturalist to map the planetary food web.

Article Title: How amateur naturalists are helping scientists track what everything eats

Article References: How amateur naturalists are helping scientists track what everything eats. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: citizen science, food web, iNaturalist, trophic interactions, ecology, biodiversity, Who Eats Whom, open-source database, PLOS Biology, public engagement, predator-prey relationships, data visualization

The project, described in a new article in the Journal of Agromedicine, began with a practical goal: giving Extension educators a way to learn more about agricultural injuries in their own communities and to compare patterns across different regions. Salah Issa, an Illinois Extension specialist and associate professor in the Department of Agricultural and Biological Engineering, part of the College of Agricultural, Consumer and Environmental Sciences and The Grainger College of Engineering, explained that the team eventually broadened the scope to include search options tailored to researchers, whose data demands differ considerably from those of frontline educators.

The result is the Agricultural Injury Dashboard, a web-based platform covering agricultural injuries in five Midwestern states: Iowa, Illinois, Michigan, Minnesota, and Wisconsin. The dashboard contains tracks for researchers and educators to customize their queries according to specific purposes, allowing users to search across a variety of parameters and view the results as interactive graphs and charts. Rather than presenting static tables or downloadable spreadsheets, the platform responds dynamically to user input, making it possible to explore the underlying data from multiple angles without any specialized analytical training.

Lead author Sihan Li, who helped develop the dashboard as a doctoral student in Agricultural and Biological Engineering and now continues the work as a visiting Extension outreach associate, emphasized the interactivity at the heart of the design. Everything on the platform responds to selection: clicking on a given year and a certain type of injury causes every chart to highlight that specific injury type and display its associated age, gender, and fatality distribution. Because the charts work in tandem, users can progressively hone in on the data according to their needs, narrowing from broad regional trends down to the specific demographic and circumstantial details of particular incident categories.

The development process began with a systematic review of existing sources where users could look up agricultural injury information. The researchers found that these resources were not very user-friendly, a shortcoming that limited their usefulness for the very audiences—educators, safety specialists, and community planners—who most needed timely and accessible data. The team set out to build something functional, efficient, and easy to use, and they refined the design through pilot testing with different user groups, incorporating feedback at each stage to improve navigation, clarity, and the relevance of the available filters.

To populate the dashboard, the researchers drew on the AgInjuryNews database, a resource that categorizes agricultural injuries reported in news coverage across the United States and Canada. For the initial release, they included AgInjuryNews data from the five Midwestern states spanning 2016 to 2024. The dataset provides rich contextual information for each incident, including the cause and type of injury, the location, the time of occurrence, and victim demographics such as age and gender. In addition, each record carries classification codes from the Occupational Injury and Illness Classification System, known as OIICS, and the Farm and Agricultural Injury Classification, or FAIC, giving the platform a standardized structure that supports rigorous comparison across years, states, and injury categories.

Li and Issa built the interactive dashboard using Microsoft Power BI, a business analytics platform that provides a wide range of data visualization capabilities. The choice allowed the team to link multiple visualizations so that they update in concert, a technical feature that underpins the dashboard’s ability to highlight selected injury types across every chart simultaneously. The partnership with AgInjuryNews supplied the initial data foundation, but the researchers plan for future updates to be populated by their own data collection efforts, giving the University of Illinois team direct control over the quality and timeliness of the records feeding the platform.

That in-house data operation is already substantial. According to Issa, the University of Illinois database for agricultural injuries is based on news clippings, and the team also asks the Illinois Department of Public Health to review the compiled information so that any missing cases can be added. He described the resulting dataset as probably one of the best available for fatal agricultural injuries in the state. While this means the dashboard will be primarily focused on Illinois going forward, Issa noted that users from other states should find it easy to adapt the platform to their own purposes by adding their own datasets, an approach that could extend the tool’s benefits well beyond the Midwest.

The researchers argue that access to a resource providing an overview of agricultural injuries is essential to education and prevention. Because accidents, especially fatal ones, happen rarely enough within any single community that patterns may go unnoticed, Issa stressed the importance of staying abreast of what is happening across the wider region. When trends of incidents become visible across multiple communities, Extension services and safety organizations can launch targeted programs to educate farmers and address emerging hazards before they grow into larger problems. In this way, the dashboard functions not merely as a record of past events but as an early-warning system that transforms scattered incident reports into actionable regional intelligence.

Looking ahead, the team hopes to expand the dashboard so that it captures not just the occurrence of accidents but also their monetary and human impact on communities, adding an economic and social dimension to the existing injury data. The research article, titled “Design of an Interactive Agricultural Injury Dashboard: Enhancing Data Visualization and Stakeholder Engagement Using AgInjuryNews Data,” is published in the Journal of Agromedicine. The project was partially supported by the Centers for Disease Control and Prevention’s National Institute for Occupational Safety and Health, with additional research support provided through the Hatch Act of 1887 from the U.S. Department of Agriculture’s National Institute of Food and Agriculture. The work also credits AgInjuryNews.org for providing access to the agricultural injury data on which the platform’s initial dataset was built, a collaboration that illustrates how open data resources and university research teams can combine to produce tools with immediate practical value for rural safety.

Subject of Research: Development of an interactive dashboard for visualizing and analyzing agricultural injury data to support safety education and prevention.

Article Title: U of I researchers develop interactive tool to track agricultural injuries, strengthen prevention efforts

Article References: U of I researchers develop interactive tool to track agricultural injuries, strengthen prevention efforts. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: agricultural safety, agricultural injuries, data visualization, University of Illinois, AgInjuryNews, Extension education, farm fatalities, Power BI, Journal of Agromedicine, injury prevention, Midwest agriculture, occupational safety

The work, published open access in the Journal of Ambient Intelligence and Humanized Computing, responds to a long-standing problem in environmental communication. Conventional charts are excellent for analysis, but research on psychological distance shows that large-scale, slow-moving phenomena such as climate change often fail to register as personally urgent when they are presented as abstract numbers. Eco-visualization, a practice first articulated in museum contexts and later explored in energy-feedback research, attempts to close that gap by blending data representation with aesthetic, narrative, and experiential elements. The catch has always been skill: designing these hybrid images demands both creative sensibility and technical competence in data processing, segmentation, and color mapping, a combination that excludes many talented designers.

The team’s answer is a structured, six-step Human-AI co-creative pipeline. The process begins with generative image creation: the user describes a dataset topic, and a text-to-image model proposes visually meaningful scenarios, such as a flooded coastal city for sea-level data or a bleached coral reef for ocean acidification. The user can accept a suggestion, request variations, or start over, iteratively steering the AI toward a contextually appropriate scene. Next comes dataset configuration, in which the user uploads structured quantitative data in CSV or TSV format; the system parses the file, displays columns as interactive chips, and pre-computes minimum and maximum values that will later define the color scale.

The third step is where the machine-learning engineering becomes visible. The system integrates Meta AI’s Segment Anything Model, or SAM, to automatically analyze the generated image and produce pixel-level masks for each identifiable object or region. Three Vision Transformer backbones of increasing size, ViT-B, ViT-L, and ViT-H, are available, but the interface deliberately hides that choice: ViT-B runs by default, and larger variants sit behind an optional advanced control. If the automatic masks miss something, a lasso tool lets users draw custom regions, and a mask editor supports merging segments through Boolean union operations, undoing recent changes, or resetting to the original output. The design philosophy is that users should evaluate the masks that matter to their visualization, not choose model architectures.

Color is treated with scientific seriousness. The system encourages palettes validated for perceptual uniformity and accessibility, such as the scientific color maps developed by geoscientist Fabio Crameri and colleagues, whose 2020 Nature Communications paper documented how arbitrary gradients can mislead viewers and distort quantitative interpretation. Users can also build custom palettes from start and end colors, and the AI suggests alternatives derived from the dominant colors of the generated image so the visualization blends naturally into its scene. During overlay generation, the chosen palette is mapped linearly onto the dataset’s value range, and the gradient is applied only inside the masked region using alpha blending, with independent hue, saturation, and lightness coefficients giving designers fine-grained control over how strongly data colors replace the original pixels. A separate edge-smoothing parameter softens boundaries between modified and untouched areas.

To understand whether this pipeline actually works for its intended audience, the researchers ran a remote user study with 13 participants, twelve of them from design backgrounds, chosen precisely because they combine visual expertise with limited experience in data-driven toolchains. Participants received no training, only a brief written introduction and a guided task: generate a mountain landscape, load monthly atmospheric CO2 data, merge masks if desired, colorize only the lake region with a scientific palette, and export the result. The evaluation combined the NASA Task Load Index for cognitive load, the short User Experience Questionnaire for pragmatic and hedonic quality, open-ended feedback, and a demographic section, all collected anonymously in compliance with GDPR.

The results were strikingly divided. Cognitive load, measured on a 0 to 100 scale, averaged 59 with a standard deviation of 16.1, squarely in the high range, and the spread suggests the mental burden varied considerably from person to person. Pragmatic quality, the UEQ-S dimension capturing usability and task effectiveness, scored a mean of minus 0.96, below the minus 0.8 threshold that marks a problematic rating. Items probing whether the system felt easy and clear scored poorly, and participants reported that the browser froze during segmentation, a consequence of running the model locally on CPU. One participant described the HSL adjustments as monotonous and struggled to grasp what the parameters actually changed, hinting that the interface failed to communicate the rationale behind its controls.

Yet hedonic quality told the opposite story, averaging 1.33, a clearly positive evaluation. Participants rated the system as highly inventive and on the leading edge, with scores of 6.23 and 5.69 respectively on those items, indicating that the core concept of embedding environmental data inside emotionally resonant imagery resonated strongly even as the interaction stumbled. The overall UEQ-S score of 0.18 landed in the neutral range, with the system’s conceptual appeal only partially compensating for its usability problems. The authors argue this gap exposes a fundamental tension in Human-AI co-creation: granting users expressive control increases creative freedom but inflates cognitive demand when the purpose of each parameter is not immediately obvious, pointing toward design strategies like progressive disclosure and contextual explanation.

The team is candid about the study’s limits. The sample was small, deliberately composed of designers, and constrained to a prescribed task with a shared dataset, so the findings speak to first-use learnability rather than open-ended creative practice. CPU-based segmentation introduced latency that degraded perceived responsiveness, and the AI’s role, while framed as collaborative, remains fundamentally reactive: it responds to prompts but does not learn from users across sessions or proactively shape the creative process, placing the current system closer to AI-assisted design than to fully bidirectional co-creation. Future work will move segmentation to GPU-accelerated or cloud inference, add contextual guidance for parameter controls, and run ecologically valid studies in which participants choose their own datasets, audiences, and visual narratives, ideally involving environmental scientists, educators, and communication professionals alongside designers.

Even with those caveats, the contribution is concrete and timely. At a moment when generative models are flooding the internet with images, this research asks a sharper question: can those models help make invisible environmental processes visible without sacrificing scientific integrity? By pairing perceptually reliable color science with iterative human steering, the pipeline sketches a workflow in which data and imagination reinforce rather than compete with each other. The mountain lake that darkens as carbon dioxide climbs is not a replacement for the chart behind it, but it is a complement that speaks to a different part of the mind, and the study’s mixed verdict, captivating concept, demanding interface, maps the road that co-creative tools must travel before that vision becomes routine practice in climate communication.

Subject of Research: A human-AI co-creative pipeline for designing data-driven eco-visualizations

Article Title: Co-creating eco-visualizations: a human–AI collaborative pipeline for data-driven visual design

Article References: Ceccarini, C., Bottari, B., Liçaj, A., Matteucci, E., & Delnevo, G. (2026). Co-creating eco-visualizations: a human–AI collaborative pipeline for data-driven visual design. Journal of Ambient Intelligence and Humanized Computing. https://doi.org/10.1007/s12652-026-05128-w

Image Credits: AI Generated

DOI: 10.1007/s12652-026-05128-w

Keywords: eco-visualization, human-AI co-creation, data visualization, generative AI, image segmentation, Segment Anything Model, environmental communication, cognitive load, user experience, scientific color palettes, climate data, co-creative systems