MakerLens : What Sign-In , Reservation and Training Data Can ( and Cannot ) Tell You
MakerLens : What Sign-In , Reservation and Training Data Can ( and Cannot ) Tell You
复制标题
MakerLens：登录、预订和培训数据可以（和不能）告诉您什么
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Nathan Khuu
中科院分区：
文献类型：
--
作者：
Nathan Khuu
数据可以帮助航天工作人员和领导层了解其航天的“脉搏”，并为战略规划和决策提供信息。在向利益相关者和资助者构思令人信服的叙述时，数据也很有用。ISAM之前的多篇论文都强调了收集和分析有关makerspace使用情况的数据的重要性。以前的工作往往遵循两种主要方法中的一种。首先，自动收集的使用或登录数据的描述性统计数据可能会显示活动模式的细粒度视图。其次，makerspace用户完成的调查可以揭示他们使用makerspace的动机和原因。在这篇文章中，我们描述了如何通过应用多个学期的聚合时间序列分析来从自动收集的makerspace数据中获得额外的见解。我们专注于makerspace中一些最普遍和最可访问的数据：(1)用户进入空间或开始使用特定机器时的登录数据；(2)流行机器的预订数据(例如，日历注册)；以及(3)个别机器类型的培训记录。许多makerspace已经为访问控制和调度目的收集了此类数据。我们展示了如何对这些数据进行几天、几周、几个月和一个学期的聚合时间序列分析，从而产生比瞬时统计数据更丰富的图景。我们这些分析的数据集是在加州大学伯克利分校雅各布斯设计创新研究所和Citris发明实验室三个学期收集的数据。雅各布斯厅的MakerSpace分为S-3、A-4、U-3、F-4、M-3[1]，每学期服务约1,000名独特的学生。关于这个Makerspace的全面介绍可以在[2]中找到。Citris发明实验室是一个卫星制造空间，与雅各布斯·霍尔共享培训和访问控制，每学期为大约350名独特的学生(S-3，A-4，U-2，F-1，M3)提供服务。当我们分析来自这些特定makerspace的数据时，我们的目标是表明我们的分析可以在其他类似的makerspace上复制。
Data can help makerspace staff and leadership understand the “pulse” of their space and inform strategic planning and decision making. Data can also be useful in crafting compelling narratives to stakeholders and funders. Multiple prior papers at ISAM have stressed the importance of collecting and analyzing data about makerspace usage. Prior work tends to follow one of two primary methodological approaches. First, descriptive statistics of automatically collected usage or sign-in data may reveal a fine-grained view of activity patterns. Second, surveys completed by makerspace users can unravel the motivations and reasons for makerspace use. In this paper, we describe how to gain additional insights from automatically collected makerspace data by applying aggregated time-series analytics over multiple semesters. We focus on some of the most pervasive and accessible data in makerspaces: (1) sign-in data when users enter a space or start using a particular machine; (2) reservation data (e.g., calendar sign-ups) for popular machines; and (3) training records for individual machine types. Many makerspaces already collect such data for access control and scheduling purposes. We show how aggregated time-series analyses of such data over days, weeks, months, and semesters can yield a richer picture than instantaneous statistics. Our dataset for these analyses is data collected over three semesters at the Jacobs Institute for Design Innovation and the Citris Invention Lab at UC Berkeley. The makerspace in Jacobs Hall is classified as S-3, A-4, U-3, F-4, M-3 [1], serving roughly 1,000 unique students per semester. A comprehensive introduction to this makerspace can be found in [2]. The Citris Invention Lab is a satellite makerspace which shares training and access control with Jacobs Hall, serving approximately 350 unique students per semester (S-3, A-4, U-2, F-1, M3). While we analyze the data from these particular makerspaces, our goal is to show that our analyses may be replicated at other similar makerspaces.