农业大数据学报 ›› 2026, Vol. 8 ›› Issue (2): 174-182.doi: 10.19788/j.issn.2096-6369.000140

• 数据应用 • 上一篇    下一篇

基于多模块协同的动物疫病数据综合分析与预警平台(ReEpi)的设计与开发

刘思焱1,2,3(), 魏丽丽1,2,3, 王靖飞1,2,3,*()   

  1. 1 中国农业科学院哈尔滨兽医研究所哈尔滨 150069
    2 动物疫病防控全国重点实验室哈尔滨 150069
    3 农业农村部动物疫病野外科学观测研究数据中心哈尔滨 150069
  • 收稿日期:2025-11-17 接受日期:2026-03-26 出版日期:2026-06-26 发布日期:2026-06-26
  • 通讯作者: 王靖飞,E-mail: wangjingfei@caas.cn
  • 作者简介:刘思焱,E-mail: 82101235755@caas.cn
  • 基金资助:
    国家重点研发计划(2023YFC2605001);国家自然基金面上项目(32272980);黑龙江省自然基金重点项目(ZD2024C005)

Design and Development of an Integrated Analysis and Early Warning Platform for Animal Epidemic Data Based on Multi-Module Collaboration

LIU SiYan1,2,3(), WEI LiLi1,2,3, WANG JingFei1,2,3,*()   

  1. 1 Harbin Veterinary Research Institute, Chinese Academy of Agricultural Sciences, Harbin 150069, China
    2 State Key Laboratory for Animal Disease Control and Prevention, Harbin 150069, China
    3 Ministry of Agriculture and Rural Affairs, Data Center for Field Scientific Observation and Research on Animal Diseases, Harbin 150069, China
  • Received:2025-11-17 Accepted:2026-03-26 Published:2026-06-26 Online:2026-06-26

摘要:

针对动物疫病数据多源异构、时空关联及小样本不平衡等特征,本文构建并实现了基于多模块协同的动物疫病智能分析与预警平台(ReEpi)。该平台覆盖“数据—分析—预警—可视化”全流程,集成数据治理、统计分析、流行病学建模、分子演化分析、空间地理分析、智能诊断及风险预警等功能模块,采用可复用的松耦合架构设计,结合Streamlit前端与Python科学计算生态,兼顾易用性与可扩展性。集成测试与初步应用显示,平台在准确性、稳定性及交互体验上表现优异,可有效支撑农业大数据环境下的动物疫病分析与辅助决策。

关键词: 农业大数据, 动物疫病, 流行病学, 时空分析, 风险预警, 可视化

Abstract:

Given the characteristics of animal disease data, such as multi-source heterogeneity, spatiotemporal correlation, and small-sample imbalance, this study constructs and implements a multi-module collaborative intelligent analysis and early warning platform for animal diseases (ReEpi). Covering the entire process of "data-analysis-early warning-visualization", the platform integrates functional modules including data governance, statistical analysis, epidemiological modeling, molecular evolution analysis, spatial geographic analysis, intelligent diagnosis, and risk early warning. It adopts a reusable and loosely coupled architecture, combined with the Streamlit frontend and Python scientific computing ecosystem, ensuring both usability and scalability. Integrated tests and preliminary applications show that the platform performs well in accuracy, stability, and interactive experience, and can effectively support animal disease analysis and auxiliary decision-making in the context of agricultural big data.

Key words: big data analytics in agriculture, animal infectious disease, epidemiology, spatial analysis, risk early warning, visualization