认为您需要仪表板？您应该改建一个笔记本。 (Think you need a Dashboard? You should build a Notebook instead.)

After first establishing themselves as a key component of the standard Business Intelligence model during the first years of the millennium, dashboards were rapidly adopted by most companies as the go-to tool to present data-driven insights and indicators.

在千禧年的最初几年中，仪表板首先成为标准商业智能模型的关键组成部分之后，仪表板被大多数公司Swift采用，成为呈现数据驱动的见解和指标的必备工具。

When Hadoop was introduced afterwards in 2007, its launch was followed by a set of Big Data technologies that radically changed how things are done behind the curtains. They allowed parallelism on a previously unimaginable scale. These changes were, for a long period, limited to data storage and data processing. Changing the way the end users accessed data felt like an unnecessary step, because dashboards were still doing a fine job.

Hadoop于2007年推出之后，其发布之后是一系列大数据技术，这些技术从根本上改变了幕后工作方式。他们允许以前所未有的规模进行并行处理。长期以来，这些更改仅限于数据存储和数据处理。改变最终用户访问数据的方式似乎是不必要的步骤，因为仪表板仍然做得很好。

In a Big Data era that completely changed how companies process their data, dashboards managed to remain the de facto standard for making sense of the mind-boggling amounts of data being produced on a daily basis. Most companies offering dashboarding solutions rapidly adapted their products to Big Data technologies. They also offered connectors that allowed dashboards to remain the undisputed go-to tool when it comes to understanding data.

在彻底改变公司处理数据方式的大数据时代，仪表板设法保持了事实上的标准，以使每天产生的数据量令人难以置信。提供仪表盘解决方案的大多数公司都将其产品快速适应了大数据技术。他们还提供了连接器，使仪表板在理解数据时仍然是无可争议的首选工具。

But with continuous changes and improvements to the standard Big Data technologies happening at a staggering pace, maybe it’s time to update the Big Data User Experience?

但是随着标准大数据技术的不断变化和改进以惊人的速度发生，也许是时候更新大数据用户体验了吗？

仪表板的问题：您总是落后一步 (The problem with dashboards: you’re always one step behind)

When they started being integrated into technology stacks at the turn of the century, dashboards answered to a clear and coherent need: presenting KPIs and data-driven insights that offer answers to established questions. They were the portal to the company’s data, and allowed people with multiple roles and needs to understand what the data has to say. In essence, dashboards were first introduced to democratize data discovery.

当它们在世纪之交开始被集成到技术栈中时，仪表板就满足了明确而协调的需求：提供KPI和数据驱动的见解，从而为既定问题提供答案。它们是公司数据的门户，并允许具有多种角色和需求的人员了解数据的含义。本质上，首先引入了仪表板以使数据发现民主化。

But at the turn of the century, data flows were very structured, the data didn’t have that much to say, and the range of questions to ask it was limited.

但是在世纪之交，数据流已经非常结构化，数据没有太多话要说，要问的问题范围也很有限。

That no longer is the case. With the exponential growth of the data being produced daily, the value of this new black gold reaches new highs every day. The volumes of data available for exploitation in this Big Data era don’t just offer answers to a specific set of questions. They offer you questions you still haven’t thought about asking yet. This led to the rise of data exploration, with data scientists trying to extract as much value from data as possible.

情况不再如此。随着每天生成的数据呈指数增长，这种新的黑金的价值每天都达到新的高点。在这个大数据时代，可用于开发的数据量不仅为特定问题提供了答案。他们向您提供您尚未想到的问题。这导致数据探索的兴起，数据科学家试图从数据中提取尽可能多的价值。

Relying on dashboards to visualize and extract value from your data means that you have to use another technology (usually notebooks) to explore it and decide what gets to be accessible through your dashboards. Such a mechanism means that the dashboard comes always at a second phase of extracting value from data. In this era where the amounts of data available allow for an infinite number of possibilities when it comes to data exploration, no dashboard could be enough to extract all of the value your data offers.

依靠仪表板来可视化数据并从数据中提取价值意味着您必须使用另一种技术(通常是笔记本电脑 )来探索它并确定可通过仪表板访问的内容。这种机制意味着仪表板始终处于从数据提取价值的第二阶段。在这个时代，可用数据量为数据探索提供了无限的可能性，没有任何仪表板足以提取数据所提供的所有价值。

Working with this two-step mechanism means that collaboration between different roles remains limited. This is because the data architectures become too complex due to the number of technologies used by the different data specialists.

使用此两步机制意味着不同角色之间的协作仍然受到限制。这是因为由于不同数据专家使用的技术数量众多，因此数据体系结构变得过于复杂。

This chain of people using different technologies for different needs means that in order to add certain insights to a dashboard, a data analyst needs to wait for a data scientist to work on the data via a notebook. In turn the data scientist may need to wait for a data engineer to offer the data in a certain structure through a script. And remember — throughout this whole time-consuming process, the value of the data keeps decreasing.

使用不同技术满足不同需求的人员链意味着，为了向仪表板添加某些见解，数据分析师需要等待数据科学家通过笔记本来处理数据。反过来，数据科学家可能需要等待数据工程师通过脚本以某种结构提供数据。请记住，在整个耗时的过程中，数据的价值一直在下降。

Multiple dashboard-providers have tried to integrate data exploration capabilities within their platforms, with Tableau notably offering an impressive Spark connector that allows you to run Spark SQL jobs directly from your dashboard. Still, the capabilities remain limited and the interactivity is only partial, which leaves the end-user always one step behind.

多个仪表板提供者已尝试将数据探索功能集成到其平台中，Tableau尤其提供了令人印象深刻的Spark连接器，该连接器使您可以直接从仪表板运行Spark SQL作业。尽管如此，功能仍然有限，并且交互性仅是部分的，这使最终用户始终落后一步。

Whether you’re using Kibana, Tableau, or Qlikview, your dashboard can offer valuable insights regarding your data. The problem with such technologies is that they were built with data discovery in mind. And because of that they neglect one key element made possible on a massive scale in this Big Data era: data exploration.

无论您使用的是Kibana，Tableau还是Qlikview，仪表板都可以提供有关数据的宝贵见解。此类技术的问题在于它们在构建时就考虑了数据发现。因此，它们忽略了在大数据时代大规模实现的一个关键要素： 数据探索 。

As data flows keep growing exponentially, dedicating the main portal to your data merely to insights means that you’re only reading the first page of a very interesting book.

随着数据流呈指数级增长，将主要门户数据仅用于洞察力意味着您仅阅读一本非常有趣的书的第一页。

笔记本，以及它们如何将交互性提高到一个全新的水平 (Notebooks, and how they take interactivity to a completely new level)

As mentioned above, notebooks have been the standard tool for data exploration for the past few years. Since the release of project Jupyter in 2014, and through the set of functionalities it offered on top of what was already available via IPython, notebooks attracted data scientists as an ideal data exploration tool thanks mainly to one key concept: interactivity.

如上所述，在过去的几年中，笔记本电脑一直是数据探索的标准工具。自2014年发布Jupyter项目以来，笔记本计算机凭借其在IPython已有功能之上提供的功能集，主要由于一个关键概念： 交互性 ，吸引了数据科学家作为理想的数据探索工具。

Thanks to kernels (within the Jupyter ecosystem) and interpreters (within Apache Zeppelin), notebooks let you explore your data through a multitude of Big Data processing technologies. They then offer immediate access to the data via built-in visualization modules and output mechanisms. Gathering both of these capabilities into the same tool is the key to using such tool for both data discovery and exploration.

借助内核(在Jupyter生态系统内)和解释器(在Apache Zeppelin中)，笔记本使您可以通过多种大数据处理技术来探索数据。然后，他们可以通过内置的可视化模块和输出机制立即访问数据。将这两种功能整合到同一个工具中，是将此类工具用于数据发现和探索的关键。

Notebooks are not only a tool that allows for direct access to data, they do so while maintaining complete interactivity. They completely blur the line that separates data scientists and data analysts and allow people with these two roles to collaborate together seamlessly.

笔记本电脑不仅是一种可以直接访问数据的工具，而且还可以保持完全的交互性。它们完全模糊了区分数据科学家和数据分析师的界限，并允许具有这两个角色的人们无缝地协作。

This works perfectly thanks to the powerful protocol that notebooks rely on and to their main building block, cells (paragraphs in Zeppelin). By offering multiple cell types (for code and text), notebooks allow for efficient collaboration.

由于笔记本电脑所依赖的强大协议及其主要构造单元-细胞(齐柏林飞艇中的段落)，它可以完美地工作。通过提供多种单元格类型(用于代码和文本)，笔记本电脑可实现高效的协作。

To show their efficiency compared to dashboards, let’s go back to the scenario we talked about earlier. In a notebook-based architecture, when a data analyst needs certain insights within a notebook, the data engineer can add a code cell within which they manipulate the data through the adequate data processing technology. Then the data scientist uses this data in another code cell to extract the desired information and offer the output to the data analyst. This all happens without any of these three data specialists leaving the notebook.

为了显示它们与仪表盘相比的效率，让我们回到前面讨论的场景。在基于笔记本的体系结构中，当数据分析人员需要在笔记本中提供某些洞察力时，数据工程师可以添加一个代码单元，在其中通过适当的数据处理技术来操纵数据。然后，数据科学家在另一个代码单元中使用此数据来提取所需的信息，并将输出提供给数据分析人员。这一切都是在这三位数据专家都没有离开笔记本的情况下发生的。

In an era where Fast Data is the norm, extracting value from your data through a structured pipeline using different tools for each step is no longer a sustainable pattern. The data that comes through an unstructured real-time data flow may offer valuable insights when used for batch processes. But it offers even more value when it’s progressively analyzed via near-real-time processing and interactive dashboards (i.e. notebooks) that offer complete access to the raw data and sophisticated visualizations.

在以快速数据为准则的时代，通过结构化的管道为每个步骤使用不同的工具从数据中提取价值已不再是可持续的模式。当用于批处理时，通过非结构化实时数据流获得的数据可能会提供有价值的见解。但是，当通过近实时处理和交互式仪表板(即笔记本电脑)进行逐步分析时，它可以提供更大的价值，这些仪表板可以完全访问原始数据和复杂的可视化效果。