Devika Jain

Harvard University

Unclaimed · AI-structured profileORCID OpenAlex

Harvard researcher linking geospatial metadata, knowledge graphs, and governance in AI-ready research data infrastructure

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-16) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

No publication-volume metric is available in this snapshot, so output is unscored. This axis would measure total papers, preprints, and venue records over time; once indexed, it will show whether the three 2026 arXiv preprints mark a broader publishing record.

Impact—

Citation counts and field-weighted influence are not yet available, so impact is unscored. This axis would measure citations, h-index-style indicators, and uptake of the reviewed preprints; as 2026 arXiv work, citation signals may take time to accumulate.

Focus0/100

Focus scores 0/100 because the field distribution is split exactly one-third each across computational social science (GeoAI, knowledge graphs), GIScience and AI governance, and information science (research data infrastructure). The three classifier labels are, however, thematically adjacent — all three papers revolve around geospatial metadata, knowledge graphs, and data governance — so a zero score may understate topical coherence.

Collaboration28/100

Collaboration scores 28/100, computed from a small network: 3 distinct coauthors appear in the raw snapshot. Recurring names are Danny EBanks, Maya Subramanian, and Daniel Ebanks; the first and third look like name variants of one person, suggesting the true pool is still tiny. All three catalog papers list Jain as last author, consistent with a senior or supervisory role on small teams.

Activity—

Temporal productivity data beyond the three 2026 preprints is unavailable, so activity is unscored. This axis would measure publication cadence and recent-year output; the snapshot covers only one publication year, which is too narrow to estimate a rhythm.

Reproducibility32/100

Reproducibility averages 32/100 across the 3 papers Evolve has reviewed (reproMean 32, reproPapers 3). Paper-level readiness is medium for the metadata-discoverability study — which quantifies bridging gains (58.5% to 63.2%) — and low for the Dataverse knowledge-graph mapping and the governance narrative review, whose proposals are explicitly unvalidated.

Research summary

Devika Jain is a researcher at Harvard University whose recent work sits at the intersection of research data infrastructure, GIScience, and the governance of AI-enabled geospatial analysis. The Evolve-reviewed catalog covers three 2026 arXiv preprints, in all of which Jain is the last-listed author. The papers share one through-line: geospatial metadata as connective tissue for scientific data. One study argues that geospatial metadata links datasets across disciplines roughly twice as well as keyword matching, and reports that LLM-based metadata enrichment lifts cross-discipline bridging from 58.5% to 63.2%, while flagging that its evaluation is partly self-judged. Another maps Harvard Dataverse as a 215,985-node knowledge graph and finds 42.9% of datasets carry geospatial metadata, proposing but not yet testing AI enrichment for place-based social science. A third narrative review organizes eight ethical and privacy risks of LLM-enabled GeoAI into an auditable controls architecture for autonomous GIS, candidly noting its proposals remain unvalidated. Bibliometrically, the profile is early-stage: three distinct coauthors, an even three-way field split, and a reproducibility average of 32/100 across the three reviewed papers. Citation, output-volume, and activity metrics are not available in this snapshot, so impact and productivity cannot be assessed.

Collaboration pattern

Collaboration in the snapshot is small and concentrated: 3 distinct coauthors, with Danny EBanks, Maya Subramanian, and Daniel Ebanks recurring across the catalog. 'Danny EBanks' and 'Daniel Ebanks' appear to be name variants of the same collaborator, which may deflate the distinct-coauthor count further. Jain sits in the last author position on all three papers (none flagged corresponding), a pattern typical of a supervising or senior contributor on compact teams.

Research keywords

geospatial metadatametadata enrichmentFAIR principlesknowledge graphsdataset discoverabilityGeoAIlarge language modelsautonomous GISdata governancelocation privacyentity resolutionspatial ethics

Highlights

Geospatial metadata bridges disciplines about twice as well as keywords
In a 2026 arXiv study on dataset discoverability, geospatial metadata connects datasets across scientific disciplines at roughly twice the rate of keywords, and LLM-based metadata enrichment lifts bridging from 58.5% to 63.2%; the authors' self-judged evaluation tempers how far the gains can be pushed.
Harvard Dataverse mapped as a 215,985-node knowledge graph
A computational social science paper models Harvard Dataverse as a knowledge graph of 215,985 nodes and finds 42.9% of datasets carry geospatial metadata; the mapping is solid, but the proposed AI enrichment for place-based government research remains untested.
Eight governance risks of LLM-enabled GeoAI mapped to auditable controls
A narrative review organizes the ethical and privacy risks of LLM-enabled autonomous GIS into eight risk areas and proposes an auditable controls architecture; the paper is candid that the proposals are unvalidated and the evidence base remains thin, with reproducibility rated low.

Timeline

2026Last-authored 'Geospatial AI, Dataverse Metadata, and the Study of Place-Based Government' (arXiv), mapping Harvard Dataverse as a 215,985-node knowledge graph in which 42.9% of datasets carry geospatial metadata.
2026Last-authored narrative review 'Toward Governance-Aware Autonomous GIS' (arXiv), mapping eight ethical and privacy risks of LLM-enabled GeoAI to an auditable controls architecture.
2026Last-authored 'Geospatial Metadata Improves Discoverability by Connecting Datasets Across Scientific Disciplines' (arXiv), reporting LLM enrichment lifts cross-discipline bridging from 58.5% to 63.2%.

Papers in the Evolve catalog

Frequent co-authors

Danny EBanks×1 Maya Subramanian×1 Daniel Ebanks×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Dimension scores are computed from a bibliometric snapshot; output, impact, and activity metrics were unavailable and are left unscored rather than estimated.
Scope is limited to the three Evolve-reviewed 2026 arXiv preprints; other publications, if any, fall outside this catalog and are not reflected.
No citation counts, venue acceptance records, funding, or career-history data was provided; none of these is inferred.
Profile is name-resolved from public bibliographic data (ORCID 0000-0001-7800-929X); frequent-coauthor records contain apparent name variants that may affect collaboration counts.