Abstract Geographers, ecologists, and other environmental scientists are typically required to utilise non-continuous measurements from various types of sensors as part of their research activities. However, data management and processing require advanced computer skills and specific knowledge of the measurement sensors. Here, we present the Tube Database (TubeDB), an easy-to-operate software system to archive, quality control, query, and further process time-series data in an efficient manner. Data are imported by loading any unprocessed raw formats as recorded by climate stations into TubeDB. When a user requests data, a query to an on-demand processing unit is created that builds an individual processing flow with the raw data as a source. The processing flow builds on the typical preparation steps for time-series data, such as temporal aggregation, quality checks, and the interpolation of missing values and transformations. The primary user interface is a web-based application to allow easy access, rapid visualisations of time-series within seconds, and powerful product-export functionality. We also provide an R package (rTubeDB) for seamless connection to the R programming environment. TubeDB enables data requesters to assemble individually processed sets of time-series within a web-browser without requiring sensor-specific knowledge or experience in time-series processing. This complements existing time-series databases, which miss climate data-specific processing and visualisation capabilities. As a self-contained application with an embedded database and webserver, TubeDB has both low hardware demands and low installation complexity. TubeDB enables easy access to time-series data for a wide range of non-computer scientists and will be useful for many research activities.

中文翻译：

---

TubeDB：气候站数据的按需处理数据库系统

摘要地理学家、生态学家和其他环境科学家通常需要利用来自各种类型传感器的非连续测量作为其研究活动的一部分。然而，数据管理和处理需要高级计算机技能和测量传感器的特定知识。在这里，我们展示了 Tube Database (TubeDB)，这是一个易于操作的软件系统，可以有效地存档、质量控制、查询和进一步处理时间序列数据。通过将气候站记录的任何未处理的原始格式加载到 TubeDB 中来导入数据。当用户请求数据时，会创建对按需处理单元的查询，该查询以原始数据为源构建单独的处理流程。处理流程建立在时间序列数据的典型准备步骤之上，例如时间聚合、质量检查以及缺失值和转换的插值。主要用户界面是一个基于 Web 的应用程序，可以轻松访问、在几秒钟内快速可视化时间序列，以及强大的产品导出功能。我们还提供了一个 R 包 (rTubeDB)，用于无缝连接到 R 编程环境。TubeDB 使数据请求者能够在 Web 浏览器中组装单独处理的时间序列集，而无需特定于传感器的知识或时间序列处理经验。这补充了现有的时间序列数据库，这些数据库缺少特定于气候数据的处理和可视化功能。作为具有嵌入式数据库和网络服务器的独立应用程序，TubeDB 具有低硬件要求和低安装复杂性。