Zuckerman, Laurel · Zuckerman, Laurel

462 rows × 15 colsverified

14 text · 1 numeric

This dataset contains the artworks listed in the Met Provenance Research Project (Archived 2005) and enhanced with additional information retrieved from a Wikidata query on the Accession Number. The dataset includes the following fields:(from Archive) Artist, Title, Url, Date creation, AccNum, (from Wikidata) item, itemLabel, inventory_number, creator, creatorLabel, inception, The_Met_object_ID .

Ellis, Chris

6,000 rows × 149 colsverified

101 categorical · 42 text · 6 numeric

Replication data for Figures 1-5 of Partisanship and the Social Construction of Deservingness in the United States.

Miller, Blake · Miller, Blake · Timothy Hildebrandt

41 rows × 8 colsverified

8 numeric

The code in this replication package reproduces every figure and table in the paper and Supplementary Materials from the provided analysis datasets (`data/ngo_data.csv`, `data/qiushi.csv`, plus three small aggregated series for the long-horizon descriptive figures) and one external dataset (Zhang and Pan 2019; see Data Availability). Three R scripts estimate the interrupted time series (ITS) models and generate all figures and tables; one Python script generates Figure S4. The replicator should expect `models.R` to run for approximately 24-48 hours, depending on available CPU cores and memory (it estimates ~230 OLS/logit models and HAC variance matrices on ~650,000 observations); the remaining scripts run in minutes.

Méndez, Roberto · Méndez Suárez, Roberto Xavier

495 rows × 6 colsverified

3 numeric · 3 text

Este dataset reúne los materiales de replicación del estudio “Transfuguismo fantasma: indisciplina sistemática en la Asamblea Nacional del Ecuador (2013-2025)”. Incluye notebooks de Jupyter que permiten reproducir los principales análisis: gráficos del Índice de Jones, diagramas de Sankey sobre el flujo del transfuguismo, diagramas de dispersión para la clasificación de la indisciplina y la tabla de resultados correspondiente. Además, se proporcionan archivos en formatos CSV y XLSX con la tabla de clasificación de indisciplina ya procesada, lo que facilita la verificación de hallazgos y el desarrollo de nuevas aplicaciones metodológicas.

Wright, Thorin

3,532 rows × 22 colsverified

14 numeric · 8 text

Replication materials for Wright (2026): Dynamics of Deadly Loyalist Violence in Northern Ireland

MIT Election Data + Science Lab

784,248 rows × 25 colsverified

18 text · 7 numeric

This dataset contains precinct-level returns for US House elections on November 8, 2022.

MIT Election Data + Science Lab

62,311 rows × 25 colsverified

17 text · 8 numeric

This dataset contains precinct-level returns, state by state, for state and federal elections on November 8, 2022.

Mukherjee, Suvrorup · Mukherjee, Suvrorup · Brooks, Nina · et al.

9,808 rows × 4 colsverified

4 numeric

This dataset consists of attendance records of 54 government schools under the Directorate of Education, New Delhi as well as information on school characteristics, location and presence of ambien PM2.5 matter in spatial buffers around each of these schools.

Aragon Sanchez, Fernando · Karaivanov, Alexander · Krishnaswamy, Karuna

360 rows × 10 colsverified

9 numeric · 1 text

Data and code to replicate main results in paper .

Coleman Snell · Hung, Jason · Hung, Jason · et al.

4,172 rows × 11 colsverified

8 text · 3 numeric

These are the documentation and panel datasets used to build the CORDA Democratic AI-Readiness Index Report (2025 Edition) (https://aiinsocietyhub.com/corda). The CORDA Democratic AI-Readiness Index measures the vulnerability of 27 democracies and near-democracies to AI-amplified democratic backsliding, scoring each country across five theoretical drivers of democratic erosion using 82 verified indicators. AI's societal impact, ranging from shaping the information environment to the concentration of power associated with AI development, has the potential to affect the democratic level of different countries. Without sufficient guidelines, AI will contribute to democratic backsliding rather than advancing democracies. We teamed up as CORDA research fellows under the mentorship of Coleman Snell to figure out a way of extending existing democratic health indices (e.g., V-Dem, Freedom House, EIU Democracy Index, RSF, IDEA, WJP) to capture AI threat environments. We believe these indices lack indicators capturing whether democracies are structurally equipped and prepared to govern AI before harms become irreversible. This is because, as currently constructed, these indices assume pre-AI democracy signals will remain the same after the adoption of AI. For us, traditional democracy signals become insufficient when AI introduces new mechanisms to amplify drivers of political change, including economic deprivation, information environment shocks, elite defection, state capacity erosion, and polarisation. We believe AI can influence them via job displacement, synthetic media generation, AI outpacing regulatory capacity, automated governance paralysis, and the production of AI elites. Moreover, current AI governance indices fail to capture these risks. In sum, having these concerns has led us to believe that measuring AI-related signals would give beneficiaries (including researchers, policy experts and industry professionals) useful insights for assessing global countries’ exposure and readiness for democratic AI risks. Overall, our objective is designing this AI-readiness index project, followed by building democratic AI amplified backsliding predictions. These predictions can give warning signals before harm becomes irreversible. In our methodology, we attempted to follow the guidelines of the OECD composite indicators regarding data imputation, normalisation, weighting, aggregation, and sensitivity testing. That said, we are open to discussions about our theoretical framing and the methodological decisions we made in constructing the index. It’s important to note that a combination of both AI-specific indicators and general democratic health indicators was used to serve as a proxy for a country’s exposure and readiness to democratic AI risks. While these indicators were helpful in constructing a proof of concept, we believe there is high value in producing original scores for the drivers. Our decision is driven by the limited signals for some of the democratic risks we are interested in.

Bates, Alexander S. · Bates, Alexander Shakeel · Phelps, Jasper S. · et al.

5 cols · 8.0 MB

3 numeric · 2 categorical

This repository contains the data associated with the published version of the manuscript "Distributed control circuits across a brain-and-cord connectome" (Bates, Phelps, Kim, Yang et al., Nature, 2026; open access; DOI: 10.1038/s41586-026-10735-w; https://www.nature.com/articles/s41586-026-10735-w; preprint v3: https://www.biorxiv.org/content/10.1101/2025.07.31.667571v3). The dataset represents a synapse-resolution reconstruction and annotation of the adult female Drosophila melanogaster central nervous system, spanning both the brain and ventral nerve cord (VNC). Included in this repository are: * Cell annotations: including soma locations, proofreading status, cell type assignments and information about cell functions * Neurotransmitter predictions * Connectivity matrices (v2 synapse-prediction snapshot, used in the paper; and the newer v3 snapshot for downstream work) * NBLAST similarity results between BANC neurons and neurons from existing connectomes including HemiBrain, FAFB (Full Adult Fly Brain), FANC (Female Adult Nerve Cord), MANC (Male Adult Nerve Cord) and maleCNS, as well as comparisons within BANC * L2 skeletal representations of neurons, generated using the pcg_skel tool * colorMIP images of all BANC neurons * Influence scores from defined source neuron groups (e.g., sensory), computed using the linear dynamical model described in the manuscript, plus the full all-to-all influence matrix * Aligned metadata linking data elements to the cell types and analyses presented in the BANC manuscript * Snapshot ZIPs of the BANC analysis code stack (bancr, bancpipeline, BANC-project, influencer, ConnectomeInfluenceCalculator, nat.ggplot, synister_banc, the-BANC-fly-connectome, fly_connectome_data_tutorial). Each code archive carries its own Zenodo DOI, listed in the per-archive documentation under `code/`; please cite the Zenodo DOI when referring to a specific software component. * Supplementary Information files from the paper The downloadable data is based on CAVE materialization 888, snapshotted on April 17, 2026, and provides a stable reference for the analyses and figures in the published paper. The aligned EM image data, the flat v888 segmentation, the nuclei and mitochondria segmentation layers, and a long-term archive of the per-neuron mesh layer all live on BossDB (https://bossdb.org/project/bates_phelps_kim_yang2025, DOI: 10.60533/boss-2025-941r). For convenience, this Dataverse deposit also includes pointer documents under external_links/ (banc_v888_segmentation.md, banc_nuclei_segmentation.md, banc_mito_segmentation.md, banc_neuron_meshes.md) that record both the BossDB DOI and the live GCS path for each layer, with code snippets for streaming via CloudVolume / TensorStore. The live, evolving reconstruction is browsable through FlyWire Codex (https://codex.flywire.ai/banc) and the BANC portal (https://banc.community), and is accessible programmatically through CAVE, the Connectome Annotation Versioning Engine (public datastack at https://global.daf-apis.com/info/datastack/brain_and_nerve_cord_public). The bulk-data source bucket is gs://lee-lab_brain-and-nerve-cord-fly-connectome/. For visual cross-dataset comparison, the BANC viewer at https://ng.banc.community/view shows the BANC EM image data and annotated meshes from v888 in a single Neuroglancer scene alongside registered neuron meshes from FAFB, FANC, HemiBrain, maleCNS and MANC. This makes it straightforward to compare a BANC neuron with its homologue or matched cell type in any of the other adult fly connectomes, without having to set up the cross-dataset registrations or load each dataset yourself. A brief guide on navigating our data: Neurons in our dataset were tracked using the CAVE (Connectome Annotation Versioning Engine) system; the public BANC datastack is at https://global.daf-apis.com/info/datastack/brain_and_nerve_cord_public. Neurons have a 'root_id', a 64-bit integer that identifies a unique neuron state, which changes each time a neuron is 'edited'. Each identified neuron in materialization 888 has a unique 'root_id'. To track 'neurons', we have tracked 'stable' points within neurons ('position') and the underlying fixed segmentation ID for an atomic set of segmented voxels associated with that position, i.e. a 'supervoxel_id'. A 'position' (a voxel) can give you a 'supervoxel_id' (a small, uneditable collection of voxels) and a 'supervoxel_id' can give you a 'root_id' (a large editable collection of supervoxels). Synaptic links also have their own 'id'. Columns appended with 'pt_' indicate that the given information was tracked from a point position that a user added to a CAVE table. A neuron is marked 'backbone_proofread' once its primary neurites or major microtubule-rich processes have been manually reviewed end to end; in 'backbone_proofread' neurons the overall morphology has been confirmed and is not expected to change radically with further work, although minor branches or a small number of synapses may still be refined. The lighter category 'roughly_proofread' identifies neurons that are recognisable but may still be missing larger branches (often due to local data artefacts); these are useful for cell type calls but not for fine connectivity. The v626 (preprint) and v850 (interim) root_ids are also retained as join keys, so users coming from preprint-era resources can cross-reference identities into v888. Researchers are encouraged to use these data in conjunction with the online resources on the BANC portal (https://banc.community) and FlyWire Codex (https://codex.flywire.ai/banc) for further annotation, exploration, and integration with community datasets. While this Dataverse is a stable snapshot of the data we used in the published paper, FlyWire Codex will serve more up-to-date and corrected data. The BANC is a live project. (Updated 2026-06-04, with volumetric-layer pointers.)

Peters, Brian

6,553 rows × 128 colsdeclared

122 numeric · 6 text

Replication package for 'The Demographic Development Finance Gap' by Brian Peters (2026). Documents that demographic structure shapes the source of capital available to developing economies, creating a 'scissors' pattern between official development assistance (ODA) and private capital flows. Uses a 189-country panel (1990-2024) of ODA, FDI, portfolio flows, and demographic principal components. Key findings: ODA flows toward young/low-income countries (Z₁ = -32.2, p = 0.007; low-income subsample -82.6, p = 0.009) while private capital accumulates in older/richer economies. 30 demographically at-risk ('RED') countries fall below income-matched peers by 21.6 percentage points of total external finance, with the gap driven primarily by FDI and portfolio equity shortfalls. ODA complementarity multiplier is 2.3 pp per 1 pp ODA/GDP (p = 0.002): each ODA dollar attracts additional private capital. 10 of the 30 RED windows are already closed; the remaining 20 retain open windows but face below-benchmark finance. Includes RED country case studies and scenario projections.

Damavandi, Hoorsana · Antia, Kersi · Kopalle, Praveen K.

966 rows × 71 colsdeclared

53 numeric · 16 text · 2 categorical

This package contains the replication data and code for Studies 2 to 3, as well as the code, data collection instructions, and data dictionary for Study 1. The data used for Study 1 is proprietary and protected under a Nondisclosure Agreement, and therefore cannot be uploaded.

Cohen, Maya · Haigis, Liana · Haigis, Liana · et al.

16 rows × 5 colsdeclared

4 text · 1 numeric

Supplemental figure

Parv Mittal

50 rows × 10 colsdeclared

7 numeric · 3 text

This replication dataset encompasses 50 highly controlled empirical trials benchmarking the compression of Deep Reinforcement Learning (DRL) policies from a high-fidelity PyBullet environment down to edge-ready deployment footprints. It explicitly evaluates the trade-offs between model parameter precision (32-bit floating-point versus 8-bit uniform quantization), Singular Value Decomposition (SVD) matrix truncation ranks, binary storage footprints, real-time ARM Cortex-M microcontroller simulation latencies, and total trajectory-tracking operational energy efficiency.

Djupe, Paul · Walker, Brooklyn · Walker, Brooklyn

2,958 rows × 26 colsdeclared

22 numeric · 4 text

This is the data and R code sufficient to replicate the findings in: Walker, Brooklyn and Paul A. Djupe. 2026. “Be Fruitful and Multiply? Complementarianism, Pronatalism, and Suppression of Reproductive Rights.” Journal for the Scientific Study of Religion DOI: 10.1111/jssr.70073

Rozas, Juan · Henríquez, Pablo A. · Henríquez, Pablo A. · et al.

204 rows × 11 colsdeclared

6 numeric · 5 text

This dataset documents the universe of roll-call votes cast in the two bodies in charge of drafting Chile's constitutional proposals between 2021 and 2023: the Constitutional Convention (4 July 2021 – 4 July 2022, 155 members) and the Constitutional Council (7 June 2023 – 7 November 2023, 50 members). It contains 4,738 plenary votes for the Convention and 1,181 for the Council, harmonized member-level and vote-level metadata, and ideological position estimates produced with three commonly used methods: W-NOMINATE, Bayesian IRT, and Dynamic IRT. The dataset includes all R scripts required to reproduce the estimates and figures, with fixed seeds and identification anchors for full numerical reproducibility.

Shea, Patrick

12,141 rows × 6 colsdeclared

5 numeric · 1 text

This repository contains the replication materials for *Hierarchy and the State* (Patrick Shea). The book develops a theory of international hierarchy — relationships in which a leading state (the US or China) exercises authority over subordinate states — and tests its empirical implications across four substantive domains: the measurement of hierarchy, property rights and expropriation, downstream political consequences (civil war, human rights, democracy), and Chinese hierarchy. Hierarchy is measured using item-response theory (IRT) models estimated in R/Stan. All subsequent statistical analysis is conducted in Stata. The archive is organized by chapter, with a shared main dataset (`hier_state_data.dta`) that each chapter's do file loads and augments with chapter-specific covariates where needed.

Amponsem, Maxwell · Amponsem, Maxwell Peprah · Guney, Selin · et al.

111,877 rows × 17 colsdeclared

11 numeric · 4 text · 2 categorical

The dataset enables the replication of empirical findings presented in the manuscript "Geopolitical Shocks, Delivery Networks, and Wheat Trade: Evidence from the Russia-Ukraine War." The primary dataset (wheat_clean.dta) contains monthly bilateral wheat trade flows (HS-1001) derived from the UN Comtrade database (2010-2024), collapsed and transformed for use in a structural PPML gravity framework. Users requiring the original raw trade records should query UN Comtrade directly (https://comtradeplus.un.org). The master do-file (00_master.do) calls four sub-do-files in sequence, reproducing all tables and figures in the manuscript. All empirical analysis is conducted in Stata using the PPML gravity estimator (ppmlhdfe).