Insight Miner: A Time Series Analysis Dataset for Cross-Domain Alignment with Natural Language

2025-12-11

AI Summary

Temporal analysis traditionally relies heavily on expert knowledge, hindering automation and scalability. Method: We introduce TS-Insights—the first large-scale, general-purpose time-series–language alignment dataset (100K samples)—and Insight Miner, a multimodal large language model (MLLM) for time-series understanding. Our approach innovatively integrates statistical feature engineering with GPT-4–assisted agent workflows to generate high-quality, natural-language time-series descriptions; treats raw time-series data as a native modality within the MLLM architecture; and enables end-to-end training via statistical feature extraction, multimodal instruction tuning, and cross-domain representation alignment. Contribution/Results: Extensive experiments demonstrate that Insight Miner significantly outperforms state-of-the-art models—including LLaVA and GPT-4—on time-series description generation. It delivers interpretable, generalizable, and AI-native temporal analytics capabilities, establishing a foundational framework for automated, language-guided time-series analysis.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Natural Language Processing: (Large) Language Models
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Time-series data is critical across many scientific and industrial domains, including environmental analysis, agriculture, transportation, and finance. However, mining insights from this data typically requires deep domain expertise, a process that is both time-consuming and labor-intensive. In this paper, we propose extbf{Insight Miner}, a large-scale multimodal model (LMM) designed to generate high-quality, comprehensive time-series descriptions enriched with domain-specific knowledge. To facilitate this, we introduce extbf{TS-Insights}footnote{Available at href{https://huggingface.co/datasets/zhykoties/time-series-language-alignment}{https://huggingface.co/datasets/zhykoties/time-series-language-alignment}.}, the first general-domain dataset for time series and language alignment. TS-Insights contains 100k time-series windows sampled from 20 forecasting datasets. We construct this dataset using a novel extbf{agentic workflow}, where we use statistical tools to extract features from raw time series before synthesizing them into coherent trend descriptions with GPT-4. Following instruction tuning on TS-Insights, Insight Miner outperforms state-of-the-art multimodal models, such as LLaVA citep{liu2023llava} and GPT-4, in generating time-series descriptions and insights. Our findings suggest a promising direction for leveraging LMMs in time series analysis, and serve as a foundational step toward enabling LLMs to interpret time series as a native input modality.

Problem

• Mining insights from time-series data requires deep domain expertise
• The process of extracting insights from time-series data is time-consuming and labor-intensive
• There is a lack of general-domain datasets for time series and language alignment

Innovation

• Large multimodal model generates time-series descriptions with domain knowledge
• Agentic workflow uses statistical tools and GPT-4 for dataset creation
• Instruction tuning on a general-domain alignment dataset improves performance