In Australia, as in other places in the world, we accumulate and curate Big Data – in the public and private spheres; as consumers and citizens; for profit, for proof and for the promises it holds. Yet, data are only ever a partial snapshot of the lives of humans, communities, environments and places. Data has its place, but in order to use data to effect change, we need to understand which questions to ask, which assumptions to suspend, which data to collect and how to interpret them, we need knowledge of the broader context. Taking a systems-level view of data allows us to ask questions about it as a technical and sense making apparatus, as well as about the consequences of data collection, storage and use. This article advocates for a cybernetic approach to thinking about and using data. Cybernetics can be broadly understood as the study of feedback, communication and control in a system. A cybernetic approach to Big Data means de-centring data to focus on a broader system of cultural, technological and environmental dimensions. In this approach, the system itself is the core unit of analysis, and also a vantage point or an approach through which to encounter and re/structure such systems. Employing a cybernetic approach allows us to move beyond attempts to order and reorder information, and instead see it as making implicit knowledge explicit, and then knowing that the hard work is what to do with that knowledge. Through this we are able to conceptualise more complex inputs, outputs and interactions, than simply those captured in a dataset. The insights drawn from a particular dataset or datasets, or from a particular model or algorithm (however well parameterised) never represent the whole truth, and a cybernetics framing enables practitioners to say, if that is the case, what now? A cybernetic approach suggests we should preference deeper understanding, not simply more truth.

中文翻译：

---

更多的数据是否等于更多的真相？走向数据的控制论方法

在澳大利亚，就像在世界其他地方一样，我们在公共和私人领域积累和管理大数据；作为消费者和公民；为了利润，为了证明和它所持有的承诺。然而，数据只是人类、社区、环境和地方生活的部分快照。数据有其一席之地，但为了使用数据来影响变革，我们需要了解要问哪些问题、暂停哪些假设、收集哪些数据以及如何解释它们，我们需要了解更广泛的背景。以系统级的数据视图，让我们能够提出问题，将其作为技术和意义制造工具，以及数据收集、存储和使用的后果。本文提倡采用控制论方法来思考和使用数据。控制论可以广义地理解为对系统中的反馈、通信和控制的研究。大数据的控制论方法意味着去中心化数据以关注更广泛的文化、技术和环境维度系统。在这种方法中，系统本身是分析的核心单元，也是遇到和重新/构建此类系统的有利位置或方法。采用控制论方法使我们能够超越对信息进行排序和重新排序的尝试，而是将其视为使隐含知识显式，然后知道如何处理这些知识是一项艰巨的工作。通过这种方式，我们能够概念化更复杂的输入、输出和交互，而不仅仅是在数据集中捕获的那些。从特定数据集或数据集得出的见解，或者来自特定模型或算法（无论参数化得多么好）永远不会代表全部真相，控制论框架使从业者能够说，如果是这样，现在怎么办？控制论方法表明我们应该更喜欢更深入的理解，而不仅仅是更多的真相。