Core concepts（核心概念）¶
This page discusses several core concepts that are fundamental to understanding the Quiver application:
Cards¶
Exploration and analysis in Quiver are performed through the use of cards, which can be chained together to perform complex operations. Some cards create charts or perform calculations, while others are used to manipulate your data by filtering, joining, deriving new columns, and so on.
Every card in Quiver takes zero or more inputs, and produces an output of a specific type, for example object set, time series, categorical chart, number, etc. Together, these types form Quiver’s data model, and define how cards can be chained together.
Cards can be configured in the editor panel by selecting the 
icon in the top-right corner of the card.
Canvas and graph¶
Quiver provides two view modes for building your analysis: canvas mode and graph mode.
A canvas is a page where you can display, rearrange, and resize the cards in your analysis. An analysis can contain multiple canvases.
:::callout{theme="neutral"} Unlike a Contour path, a Quiver canvas is used for display and organization only. Rearranging cards on your canvas will not affect the underlying sequence of data transformation. :::
In graph mode, cards are represented as nodes and their connections as links, using a left-to-right layout that makes it easy to follow data flow. You can organize the graph using color groups, and hide or filter nodes to focus on a specific area of your analysis.
Cards added in graph mode are not automatically placed on a canvas. You can add or remove cards from a canvas at any time.
:::callout{theme="neutral"} Graph mode is best for tabular analysis, while canvas mode is best for time series analysis. Regardless of data type, canvas mode is better for presenting final outputs such as charts and tables, and graph mode is better for inspecting card dependencies. :::
Learn more about canvas mode and graph mode.
Objects¶
In Quiver, objects from the Ontology are used as the primary data input for tabular analysis. Quiver natively supports many different cards for filtering, transforming, and visualizing objects data.
For advanced transformations on objects data, such as deriving properties and joining between linked objects, users can also leverage Quiver’s suite of transform table, materialization, and function cards.
Learn more about object analysis.
Time series¶
Quiver has first-class support for time series analysis. Time series are primarily added to Quiver through time series properties, however time series syncs can also be viewed directly.
Quiver provides an extensive library of transformations and visualizations for time series data. Quiver also supports advanced time series workflows such as anomaly detection and event analysis.
Additionally, time series transformations can be saved as derived series, or used to create alerts with time series automations.
Learn more about time series analysis.
Dashboards¶
With Quiver, you can build interactive dashboards that display the results and findings of your analyses. These dashboards can be used as standalone views, or embedded in other Foundry applications such as Workshop.
Parameters¶
Quiver parameters allow you to easily switch between different views of the data and results. After creating parameters, you can use them in your cards and expose them in your dashboard. This allows end users of a dashboard to interact live with the data and results presented.
Saving and versioning¶
Quiver analyses are saved manually by clicking the Save button in the top right of the application. A version history is also provided, allowing you to view or revert your analysis to previous saved versions. Additionally, in between each Save action, Quiver auto-saves your "working" state (storing it in the state URL variable, for example state=j05na7mun3). This allows you to refresh your page and get back your exact analysis state even if you have not saved. Note that if you are sharing a URL link with the state variable set, this will open that working state rather than the latest analysis version.
If multiple users are working on the same analysis at the same time, they are able to work independently without interference, however saving changes will overwrite each others saved changes.
Learn more about saving and versioning.
中文翻译¶
核心概念¶
本页介绍理解 Quiver 应用所需掌握的若干核心概念：
卡片¶
在 Quiver 中，探索和分析通过使用卡片（cards）来完成，卡片可以串联起来执行复杂操作。有些卡片用于创建图表或执行计算，而另一些则用于通过筛选、连接、派生新列等方式操作数据。
Quiver 中的每张卡片接收零个或多个输入，并生成特定类型的输出，例如 object set（对象集）、time series（时间序列）、categorical chart（分类图表）、number（数字）等。这些类型共同构成了 Quiver 的数据模型，并定义了卡片之间如何串联。
可以通过点击卡片右上角的 图标，在编辑面板中对卡片进行配置。
画布与图谱¶
Quiver 提供两种视图模式来构建分析：画布（canvas）模式和图谱（graph）模式。
画布是一个页面，您可以在其中显示、重新排列和调整分析中卡片的大小。一个分析可以包含多个画布。
:::callout{theme="neutral"} 与 Contour 路径不同，Quiver 画布仅用于显示和组织。在画布上重新排列卡片不会影响底层的数据转换顺序。 :::
在图谱模式下，卡片以节点（nodes）表示，其连接关系以链接（links）表示，采用从左到右的布局，便于追踪数据流。您可以使用颜色分组来组织图谱，并隐藏或筛选节点以专注于分析的特定区域。
在图谱模式下添加的卡片不会自动放置在画布上。您可以随时在画布上添加或移除卡片。
:::callout{theme="neutral"} 图谱模式最适合表格分析，而画布模式最适合时间序列分析。无论数据类型如何，画布模式更适合呈现图表和表格等最终输出，而图谱模式更适合检查卡片之间的依赖关系。 :::
了解更多关于画布模式和图谱模式的信息。
对象¶
在 Quiver 中，来自本体（Ontology）的对象被用作表格分析的主要数据输入。Quiver 原生支持多种不同的卡片，用于筛选、转换和可视化对象数据。
对于对象数据的高级转换，例如派生属性以及在链接对象之间进行连接，用户还可以利用 Quiver 的转换表（transform table）、物化（materialization）和函数（function）卡片套件。
时间序列¶
Quiver 对时间序列（time series）分析提供了一流支持。时间序列主要通过时间序列属性（time series properties）添加到 Quiver 中，但也可以直接查看时间序列同步（time series syncs）。
Quiver 为时间序列数据提供了丰富的转换（transformations）和可视化（visualizations）库。Quiver 还支持高级时间序列工作流，例如异常检测（anomaly detection）和事件分析（event analysis）。
此外，时间序列转换可以保存为派生序列（derived series），或用于通过时间序列自动化（time series automations）创建警报。
仪表盘¶
使用 Quiver，您可以构建交互式仪表盘（dashboards）来展示分析的结果和发现。这些仪表盘可以作为独立视图使用，也可以嵌入到其他 Foundry 应用中，例如 Workshop。
参数¶
Quiver 参数（parameters）允许您轻松切换数据和结果的不同视图。创建参数后，您可以在卡片中使用它们，并将其暴露在仪表盘中。这使得仪表盘的最终用户能够与所呈现的数据和结果进行实时交互。
保存与版本管理¶
Quiver 分析通过点击应用程序右上角的“保存”按钮手动保存。同时还提供版本历史记录，允许您查看或将分析恢复到之前的已保存版本。此外，在每次“保存”操作之间，Quiver 会自动保存您的“工作”状态（将其存储在 state URL 变量中，例如 state=j05na7mun3）。这允许您刷新页面并恢复精确的分析状态，即使您尚未保存。请注意，如果您分享带有 state 变量设置的 URL 链接，打开的是该工作状态，而非最新的分析版本。
如果多个用户同时处理同一个分析，他们可以独立工作而互不干扰，但保存更改会覆盖彼此的已保存更改。