Connecting analytics to operations（连接分析与运营）¶
Although Foundry's focus is upon enabling operational workflows that capture data back into the system, the results of work done using Foundry's analytical capabilities is also a key part of the puzzle. This page covers how analytical tools can be used to deliver value to end users in a cohesive way that empowers decision-making.
Dashboards¶
Often, the results of open-ended exploration of data in the Ontology can result in interesting insights that could be useful for other people in your organization. Starting with a set of objects, drilling down into a subset, and exploring properties using charts or linked objects may yield a repeatable workflow that can be combined with other platform capabilities in a valuable way.
Quiver is the application used for open-ended analysis on object data in Foundry. Quiver dashboards enable you to create parameterized, curated views that end users can use to explore data in a structured fashion. Once you've published a dashboard, you can use it in operational applications in a few different ways:
- Quiver dashboards can be embedded into Workshop applications, allowing you to flexibly show dashboards to users as part of a structured workflow.
- Quiver dashboards can be embedded into Carbon so that users can access the template as a tab in their workspace.
Using this functionality, exploratory analysis can quickly transition from an ad-hoc dashboard or insight to being a core part of an operational workflow.
Models¶
Foundry's machine learning capabilities provide a full suite of tools for model development, integration, evaluation, and deployment. Once models have been validated according to organizational criteria, they can be deployed to end users easily.
Workshop's support for Scenarios enables the deployment of machine learning directly to end users without requiring any custom development or extensive configuration. Users interact with concepts they are familiar with, while seeing model-backed results in an easy-to-understand form such as a forecast or estimate.
For more details on tying machine learning to organizational outcomes, refer to Models in the Ontology.
中文翻译¶
连接分析与运营¶
虽然 Foundry 的核心在于支持能够将数据回传至系统的运营工作流，但利用 Foundry 分析能力所完成的工作成果同样是整体解决方案的关键部分。本页将介绍如何以统一的方式运用分析工具，为最终用户创造价值，从而赋能决策制定。
仪表盘¶
通常，对本体（Ontology）中数据进行开放式探索的结果，可能会产生对组织内其他人员有用的深刻见解。从一组对象入手，深入分析子集，并利用图表或关联对象探索属性，可能会形成可重复的工作流，进而与其他平台功能相结合，产生有价值的成果。
Quiver 是用于在 Foundry 中对对象数据进行开放式分析的应用程序。Quiver 仪表盘（dashboards）允许您创建参数化、精选的视图，最终用户可通过这些视图以结构化方式探索数据。发布仪表盘后，您可以通过以下几种方式将其用于运营应用程序：
- Quiver 仪表盘可嵌入 Workshop 应用程序，从而在结构化工作流中灵活地向用户展示仪表盘。
- Quiver 仪表盘可嵌入 Carbon，使用户能够以工作区标签页的形式访问该模板。
利用此功能，探索性分析可以快速从临时仪表盘或洞察转变为运营工作流的核心组成部分。
模型¶
Foundry 的机器学习能力提供了一整套用于模型开发、集成、评估和部署的工具。一旦模型根据组织标准通过验证，便可轻松部署给最终用户。
Workshop 对场景（Scenarios）的支持使得机器学习能够直接部署给最终用户，无需任何自定义开发或大量配置。用户可以与熟悉的概念进行交互，同时以易于理解的形式（如预测或估算）查看基于模型的结果。
有关将机器学习与组织成果相结合的更多详情，请参阅本体中的模型。