Explore chapters and articles related to this topic
Data Quality and Inference Errors
Published in Ian Foster, Rayid Ghani, Ron S. Jarmin, Frauke Kreuter, Julia Lane, Big Data and Social Science, 2020
The core issue that often is the cause of these errors is that such data may not be generated from instruments and methods designed to produce valid and reliable data for scientific analysis and discovery. Rather, this is data that are being repurposed for uses not originally intended. It has been referred to as “found” data or “data exhaust” because it is generated for purposes that often do not align with those of the data analyst. In addition to inadvertent errors, there are also errors from mischief in the data generation process; for example, automated systems have been written to generate bogus content in social media that is indistinguishable from legitimate or authentic data. Social scientists using this data must be keenly aware of these limitations and should take the necessary steps to understand and hopefully mitigate the effects of hidden errors on their results.
Analyzing Big Data
Published in Stephan S. Jones, Ronald J. Kovac, Frank M. Groom, Introduction to COMMUNICATIONS TECHNOLOGIES, 2015
Stephan S. Jones, Ronald J. Kovac, Frank M. Groom
Among the multiple characteristics of big data, four stand out as its defining characteristics. A huge volume of data. The tools that store, access, and analyze the data must be able to manage billions of lines of data at one time.A complexity of data types and structures. Available and useful data are composed of an increasing volume of unstructured data. About 80%–90% of these data in which businesses, governments, and private organizations are interested are classified as unstructured data, which are sometimes described as part of the digital shadow or data exhaust of a process. Data exhaust is the data in social networks where the residue of communication, which is like the exhaust fumes of the user’s communication process, might prove to be useful when collected and offered to a marketer wishing to discover and analyze collective and individual social behaviors.Constantly increasing speed in the creation and sharing of these new unstructured data. Every day brings forth a new batch of data, merging with previous data, or obliterating already stored and analyzed data.Distinct analytical processes. Due to the size and variety of structures exhibited by these new data, they cannot be easily and efficiently analyzed using only traditional methods applied against traditional databases. Where previously we might be storing records of text and numbers in a structured relational database with its fixed table structure, we now have blobs of unstructured but interesting data, which are difficult to acquire, parse, store, access, and then analyze.
Data Philanthropy: Corporate Responsibility with Strategic Value?
Published in Information Systems Management, 2020
Jordana J. George, Jie (Kevin) Yan, Dorothy E. Leidner
An interesting aspect of RBV is that it does not focus on any particular type of resources. Resources may be human, or a process, or a unique combination of people, places, and things. An important quality of resources is dependence on context (Artz & Brush, 2000). For example, a half empty plastic water bottle would be considered trash when it is tossed in a garbage can but that bottle could save a life if it is found in a desert by a lost hiker. The value of data also depends on context. Data exhaust, for example, is the voluminous quantity of data spewing from IoT, smart city devices, and other internet enabled systems. It could be considered trash in some contexts (European Commission, 2017). But in the appropriate context, data exhaust can provide traffic patterns, predict road repair schedules, or forecast electricity blackouts. In this case, one person’s trash becomes another’s treasure. Whether data is trash or treasure comes down to the context and if it is treated as a by-product or a resource of the firm.
Crafting animated parables: an embodied approach to representing lifestyle behaviours for reflection
Published in Digital Creativity, 2021
As technologies have been increasingly integrated into every facet of daily life, more data are generated, extracted, and aggregated around individuals’ behaviours via various sources, such as electronic payment systems, IoT (Internet of Things) products, surveillance cameras, and the so-called ‘data exhaust’ (Zuboff 2015). Some of these systems present aggregated information back to individual users in typical data visualization, such as various bar charts shown in the latest iOS on the iPhone about the user’s daily and weekly screen time (in hours and minutes) spent on different mobile applications (apps). Such visualizations require deliberate reading and are not effective for daily review, because people do not always act like rational data analysts (Rooksby et al. 2014), not to mention those who are not data-savvy (Wilson, Bhamra, and Lilley 2015). General users prefer an ‘emotional link’ to the data and an ‘impressionistic image’ of the information, in other words, a representation of one’s lifestyle in an intuitive metaphor (Rapp and Cena 2016). Early HCI studies have explored representing lifestyle behavioural data in a figurative way for users’ reflection, such as Fish‘n’Steps (Lin et al. 2006), UbiFit (Consolvo et al. 2008), UbiGreen (Froehlich et al. 2009), Playful Bottle (Chiu et al. 2009), and Eco-island (Shiraishi et al. 2009). These representations typically map values of tracked data to states of on-screen virtual items like fish, flowers, trees, a garden, or an island. Findings from these studies with longitudinal field trials (from 3 weeks to 3 months) support positive correlation between virtual outcomes and participants’ emotional attachment. Using virtual environments and items to show ‘metaphorical’ outcomes of behaviours is commonly regarded as a means to complement typical data visualization (Chow, Leong, and Lee 2018; Consolvo et al. 2014; Froehlich et al. 2012; Lane et al. 2014; Pinske et al. 2012; Reitberger, Spreicer, and Fitzpatrick 2014).