Natural Language Interfaces for Spatial and Temporal Databases: A Comprehensive Overview of Methods, Taxonomy, and Future Directions

Samya Acharja; Kanchan Chowdhury

cs.DBMar 24, 2026

Natural Language Interfaces for Spatial and Temporal Databases: A Comprehensive Overview of Methods, Taxonomy, and Future Directions

Authors: Samya Acharja, Kanchan Chowdhury

Organizations: Department of Computer Science, Marquette University, Milwaukee, Wisconsin, USA

Abstract

The task of building a natural language interface to a database, known as NLIDB, has recently gained significant attention from both the database and Natural Language Processing (NLP) communities. With the proliferation of geospatial datasets driven by the rapid emergence of location-aware sensors, geospatial databases play a vital role in supporting geospatial applications. However, querying geospatial and temporal databases differs substantially from querying traditional relational databases due to the presence of geospatial topological operators and temporal operators. To bridge the gap between geospatial query languages and non-expert users, the geospatial research community has increasingly focused on developing NLIDBs for geospatial databases. Yet, existing research remains fragmented across systems, datasets, and methodological choices, making it difficult to clearly understand the landscape of existing methods, their strengths and weaknesses, and opportunities for future research. Existing surveys on NLIDBs focus on general-purpose database systems and do not treat geospatial and temporal databases as primary focus for analysis. To address this gap, this paper presents a comprehensive survey of studies on NLIDBs for geospatial and temporal databases. Specifically, we provide a detailed overview of datasets, evaluation metrics, and the taxonomy of the methods for geospatial and temporal NLIDBs, as well as a comparative analysis of the existing methods. Our survey reveals recurring trends in existing methods, substantial variation in datasets and evaluation practices, and several open challenges that continue to hinder progress in this area. Based on these findings, we identify promising directions for future research to advance natural language interfaces to geospatial and temporal databases.

Figures & tables

Panel of Figure 1 — Fig. 1 : A sample of the Business relation in Yelp, a sample natural language question with the ground truth SQL query, and 4 queries translated by various NL-2-SQL approaches as well as the output relations for such queries.

Dataset Family	Representative Datasets	Strengths	Limitations
Benchmark datasets	GeoQuery [ 56 ] , Restaurants [ 56 ] , GeoAnQu Corpus [ 22 ] , GeoSQL-Bench [ 57 , 58 ] , TRUCE [ 59 ] , SUSHI [ 59 ] , TACO [ 60 ] , NALMOBench [ 61 ] , Text-to-PromQL Benchmark [ 62 ]	Include predefined natural-language questions paired with executable queries, which makes it easier to evaluate whether a system correctly interprets spatial or temporal operators.	These datasets often do not reflect the ambiguity that appear in real-world GIS databases.
Real-world GIS datasets	Oracle Spatial Layers [ 63 ] , Denton County GIS [ 21 ] , TIGER/OSM/MODIS [ 64 ] , berlintest [ 65 , 66 ] , nanjingtest [ 65 , 66 ] , chinawater [ 66 ] , NYC GIS [ 11 ] , OSM Bavaria + SoilMap [ 67 ] , SpatialQueryQA [ 68 ] , KaggleDBQA [ 68 , 69 ] , SpatiaLite [ 12 ] , GeoQueryJP [ 69 ] , TourismQA [ 70 ]	These datasets come from real GIS databases or OpenStreetMap layers, allowing NLIDB systems to be tested in realistic spatial environments.	These datasets usually do not include predefined natural-language query sets, so researchers often need to design their own evaluation queries, which makes results difficult to compare across different NLIDB systems.
Synthetic / domain-specific datasets	BerlinMOD trajectories [ 71 ] , SECONDO trajectory demo [ 72 ] , MOQ + BT trains [ 73 ] , Extended MOQ + BT [ 74 ]	These datasets emphasize specific spatial or spatiotemporal query tasks, which helps researchers examine how NLIDB systems interpret trajectory relations, temporal predicates, and moving-object queries.	These datasets are artificially generated, often simplifying real geographic contexts, and therefore may not capture the data irregularities and interpretation challenges commonly found in real GIS databases.

TABLE I: Major dataset families used in ST-NLIDB research: representative datasets, strengths and limitations

Dataset (Short Name)	Year	Spatial	Temporal	Benchmark	RealGIS	Synthetic
Oracle Spatial Layers [ 63 ]	2007	✓	✗	✗	✓	✗
Denton County GIS [ 21 ]	2009	✓	✗	✗	✓	✗
BerlinMOD Trajectories [ 71 ]	2009	✓	✓	✓	✗	✓
TIGER / OSM / MODIS [ 64 ]	2015	✓	✗	✗	✓	✗
GeoQuery / Restaurants [ 56 ]	2019	✓	✗	✓	✗	✗
SECONDO Traj (Demo) [ 72 ]	2020	✓	✓	✗	✗	✓

TABLE II : Categories and publication years of datasets used across various ST-NLIDBs

Dataset / Data Source	Paper(s) Using Dataset	#NLQs	Query Language	NLQ Origin	Split	Public
Operational GIS Databases (Execution-Centric)
Oracle Spatial Layers [ 63 ]	[ 63 ]	N/A	Spatial SQL	None	✗	✓
Denton County GIS [ 81 ]	[ 21 ]	30	SQL	Human	✗	✓
BerlinMOD Trajectories [ 71 ]	[ 71 ]	N/A	SECONDO	None	✗	✓
SECONDO Demo Traj [ 76 ]	[ 72 ]	N/A	SECONDO	None	✗	✓
BT Berlin Trains [ 76 ]	[ 73 ] , [ 74 ]	N/A	SECONDO	None	✗	✓

TABLE III : Key characteristics of datasets across all included studies

Evaluation Dimension	Spatial NLIDB	Spatiotemporal NLIDB	Time-Series NLIDB
Execution-Based accuracy	Correct execution of spatial queries	Correct execution of time-aware queries	Not applicable
Interpretation / Translation Accuracy	Correct mapping of NL queries to spatial intent	Correct identification of spatiotemporal intent	Correct alignment of text with signal semantics
Precision-Oriented accuracy	Precision of spatial results	Precision of trajectory or temporal outputs	Precision of retrieved time-series segments
Structural / Syntactic accuracy	Optional SQL structure matching	Rarely used	Not applicable
Retrieval Quality (Ranking-Based)	Rare	Occasional	Primary (e.g., mAP)
Efficiency and Scalability	Runtime and performance	Trajectory-length and time-range efficiency	Embedding and retrieval latency

TABLE IV: Significance of various evaluation metrics across spatial, spatiotemporal, and time-series NLIDB systems

Paper	Year	Model Type	DBMS	Evaluation Metric (Normalized)	NLIDB Category
[ 92 ]	2000	Fuzzy grammar + possibility theory-based semantic parsing	Relational GIS (SQL preprocessor)	Qualitative / Human Evaluation	Spatial NLIDB
[ 63 ]	2007	Menu-based NLI + rule-based grammar	Oracle Spatial	Qualitative / Human Evaluation	Spatial NLIDB
[ 21 ]	2009	Rule-based NLP + semantic parsing	PostGIS	Execution-Based Accuracy	Spatial NLIDB
[ 71 ]	2009	Data generation and benchmark framework	SECONDO	Efficiency and Scalability	spatiotemporal NLIDB
[ 64 ]	2015	MapReduce spatial framework (non-NLI)	HDFS / SpatialHadoop	Efficiency and Scalability	Spatial NLIDB (Context)
[ 56 ]	2019	Seq2Seq GRU + spatial comprehension model	Custom geospatial DB	Interpretation / Translation Accuracy	Spatial NLIDB

TABLE V : Summary of included studies with evaluation metrics and NLIDB category

Fig. 2 : Unified methodological pipeline for NL interfaces to spatial, spatiotemporal, and time-series databases

Pipeline Stage	Spatial NLIDB	Spatiotemporal / Moving Objects NLIDB	Time-Series NLIDB
Natural Language Parsing	Entity extraction focused on places, spatial relations, and numeric constraints	Entity extraction extended with temporal expressions, object identities, and trajectory references	Text understanding aligned to signal patterns, trends, or comparative descriptions
Domain Mapping	Mapping to spatial schemas, layers, and geometry types (point, line, polygon)	Mapping to spatiotemporal schemas, trajectory identifiers, regions, and temporal indexes	Alignment between text descriptions and signal representations in a shared embedding space
Query / Task Identification	Classification into spatial query types (range, NN, spatial join, aggregation)	Richer classification including interval, trajectory similarity, cross-region, and detour queries	Implicit task formulation as retrieval or comparison rather than explicit query typing
Query Construction Strategy	Template-based spatial SQL or algebraic operators	Structured spatiotemporal operators with time-aware constraints and index selection	No explicit SQL generation; retrieval via embedding similarity and contrastive learning
Execution Model	Single-shot spatial query execution	Execution with temporal filtering, trajectory operators, and optimization-aware planning	Similarity-based retrieval over learned representations, not database execution
Result Validation	Execution success and spatial result correctness	Execution accuracy combined with efficiency and trajectory correctness	Retrieval accuracy based on embedding similarity and ranking metrics

TABLE VI: Pipeline characteristics across spatial, spatiotemporal, and time-series NLIDB systems.

Method Family	Representative Approaches	Strengths	Limitations
Rule-based / Grammar-driven NLIDBs	Rule-based NLP pipelines [ 97 , 98 , 99 ] , Handcrafted grammars [ 92 , 63 , 22 ] , Template-based query generation [ 21 ]	These systems translate queries through predefined linguistic rules, making the reasoning process easier to trace and the generated queries easier to inspect.	These systems rely on handcrafted grammars and rules, which makes them difficult to scale and adapt to diverse or evolving query formulations.
Neural / Machine-learning NLIDBs	Seq2Seq models [ 1 , 2 ] , LSTM-based classifiers [ 73 , 65 , 66 ] , Neural semantic parsing [ 100 , 3 , 101 , 102 ]	These systems learn relationships between natural-language queries and database operations from training data, allowing them to handle linguistic variation more effectively than purely rule-based systems.	These systems depend on labeled training data and learned representations, which can reduce interpretability and limit generalization to unseen query types or database schemas.
LLM-based NLIDBs	Prompt-based text-to-SQL systems [ 11 , 103 , 104 ] , Schema-aware LLM prompting [ 12 , 69 , 105 , 106 ]	These systems use LLMs to interpret natural-language questions and generate database queries through prompt-based reasoning, allowing them to handle a wide range of query expressions without relying on handcrafted rules.	These systems still suffer from spatial reasoning errors and may hallucinate database tables, columns, or spatial functions that do not exist in the database schema.
Multi-agent NLIDBs	Agentic LLM pipelines [ 93 ] , Task decomposition frameworks [ 68 , 67 ] , Reasoning + tool use [ 107 , 108 , 35 , 109 ]	These systems split the query task across multiple agents, allowing complex queries to be solved through step-by-step reasoning, tool use, and verification.	Because these systems rely on multiple interacting agents, they introduce additional system complexity, increase query latency due to multiple model calls, and require higher computational cost.

TABLE VII: Method families in ST-NLIDB systems: representative techniques, strengths, and limitations

Fig. 3 : General flow structure of rule-based and grammar-driven NLIDBs

Fig. 4 : Generic pipeline of neural net and semantic parsing-based hybrid NLIDBs

Fig. 5 : A general illustration of how LLM-prompting approaches generate database queries.

Fig. 6 : A general illustration of how multi-agent approaches generate database queries.

Method Class	Interpretation Mechanism	Learned Components	Schema Understanding	Query Planning	Execution & Validation	Ambiguity Handling	Extensibility
Rule-Based & Grammar-Driven	Handcrafted grammars and rules	None	Manual, rule-based mapping	Deterministic	Built-in constraints	Explicit rule encoding	Low
Neural & Hybrid Semantic Parsing	Neural classifiers with structured templates	Partial (e.g., LSTM, embeddings)	Knowledge-base assisted	Template-driven	Deterministic execution	Learned + rule-based	Medium
LLM-Based Interfaces	Prompt-based language reasoning	Fully learned	Prompt- and schema-context understanding	Generative	Limited or iterative	Implicit	High
Multi-Agent Architectures	Coordinated agent reasoning	Fully learned	Agent-mediated schema reasoning	Planned and iterative	Explicit agent validation	Explicit reasoning steps	Very High

TABLE VIII: Methodological characteristics of natural language interfaces for spatial, spatiotemporal, and time-series databases.

May 11, 2026

Much of Geospatial Web Search Is Beyond Traditional GIS

Web search queries concern place far more often than existing labelling schemes suggest, yet the landscape of geospatial web search queries - what people ask of place, and how often - remains poorly characterised at scale. We apply dense sentence embeddings, a lightweight SetFit classifier, and density-based clustering to the full MS MARCO corpus of 1.01 million real Bing queries without prior filtering for toponyms or spatial keywords, identifying 181,827 geospatial queries (18.0%), nearly threefold the 6.17% labelled as Location in the original annotations. The resulting taxonomy of 88 query categories reveals that geospatial web search is dominated by transactional and practical lookups: costs and prices alone account for 15.3% of geospatial queries, nearly twice the size of the entire physical geography theme. Much of this activity - costs, opening hours, contact details, weather, travel recommendations - falls outside the scope of what traditional GIS and knowledge graphs are built to serve. The categories vary substantially in the kind of answer they admit, from deterministic lookups answerable from spatial databases or knowledge graphs to evaluative or temporally volatile queries that require generative or real-time systems. We discuss implications for hybrid retrieval architectures and for benchmarks of geographic reasoning in large language models. We openly release the labelled dataset, classifier, and taxonomy.

Ilya Ilyankou, Stefano Cavazzi, James Haworth

SpaceTimeLab, Department of Civil, Environmental, and Geomatic Engineering, UCL, London, UK · Ordnance Survey, Southampton, UK

Information Retrieval

Ask PDF

Feb 19, 2026

Sonar-TS: Search-Then-Verify Natural Language Querying for Time Series Databases

Natural Language Querying for Time Series Databases (NLQ4TSDB) aims to assist non-expert users retrieve meaningful events, intervals, and summaries from massive temporal records. However, existing Text-to-SQL methods are not designed for continuous morphological intents such as shapes or anomalies, while time series models struggle to handle ultra-long histories. To address these challenges, we propose Sonar-TS, a neuro-symbolic framework that tackles NLQ4TSDB via a Search-Then-Verify pipeline. Analogous to active sonar, it utilizes a feature index to ping candidate windows via SQL, followed by generated Python programs to lock on and verify candidates against raw signals. To enable effective evaluation, we introduce NLQTSBench, the first large-scale benchmark designed for NLQ over TSDB-scale histories. Our experiments highlight the unique challenges within this domain and demonstrate that Sonar-TS effectively navigates complex temporal queries where traditional methods fail. This work presents the first systematic study of NLQ4TSDB, offering a general framework and evaluation standard to facilitate future research.

Zhao Tan, Yiji Zhao, Shiyu Wang +5

Jiangxi University of Finance and Economics · Griffith University · Yunnan University +2

Time Series QA Temporal IR Inference-Time Search

Ask PDF

Oct 8, 2026

GeoFitBench-LLM: A Benchmark for Assessing Large Language Models' Capabilities in Geospatial Data Requirements Elicitation for Geo-Analytic Tasks

Large language models are becoming central to GeoAI systems, yet their evaluation rarely considers the elicitation of geospatial data requirements. This omission is critical as retrieval oriented geo-analytic tasks cannot be supported reliably without first identifying the necessary datasets and constraints. GeoFitBench-LLM is introduced as a benchmark for assessing whether LLMs can infer the datasets, the supporting resources and the constraints required by geo-analytic requests in natural language. The benchmark is structured around two reasoning dimensions, Spatio-Temporal Relational and Semantic Grounding, reflecting long-standing challenges in geospatial information retrieval and related natural language tasks. Ground truth requirements are encoded in a fixed, provider-independent schema for reproducible evaluation of resource identification and attribute-level specification. Experiments on seven models under minimal prompting indicate that identifying the necessary resources is often achievable, with GPT-5.4 Thinking achieving the best result by recovering the complete set of resources in 66.67% of benchmark requests. Results further indicate that accurate fitness-for-use specification remains more difficult as inferential demands increase.

Amirhossein Vahdat, Thierry Badard, Luc Lamontagne

Département des Sciences Géomatiques, Université Laval, QC, Canada · Département d’informatique et de génie logiciel, Université Laval, QC, Canada

LLM Evaluation Geospatial Reasoning Benchmarks Geospatial Reasoning LLM Grounding Requirements Engineering Spatial Reasoning Benchmarks

Ask PDF

Abstract

Explore similar work

Much of Geospatial Web Search Is Beyond Traditional GIS

Sonar-TS: Search-Then-Verify Natural Language Querying for Time Series Databases

GeoFitBench-LLM: A Benchmark for Assessing Large Language Models' Capabilities in Geospatial Data Requirements Elicitation for Geo-Analytic Tasks