Data Innovation Lens: A New Way to Approach Data Design as Value Creation

Presentation at the London School of Economics and Political Science on May 10, KIN Center for Digital Innovation, Amsterdam on May 7, and at ESSEC Business School, Paris on April 30, 2024 on the study of data as innovation. The presentation is based on a paper coauthored with Marta Stelmaszak.

1. May 10, 2024
The London School of Economics and Political Science
Data Innovation Lens:
A New Way to Approach Data Design as Value Creation
Aleksi Aaltonen
Temple UniversityMarta Stelmaszak
Portland State University
WORK-IN-
PROGRESS

2. PhD 2012
Moves, acquired
by Facebook 2014
Assistant Professor
2014–2018
Assistant Professor
2018–

3. Overview
1.Motivating and positioning the study of data design
2.Data Innovation Lens: Studying data design as innovation
3.Case Vignette: Social justice through data innovation
4.Research Agenda: Toward contributions to established streams of
research and deepening the understanding of data innovation

4. Taft oil well blow
-out in Kern County, ca. 1920 (Public domain, from
https://
en.wikipedia.org
/wiki/File:Taft_oil_well_blow
-out_in_Kern_County,_ca._1920_%28CHS
-2498%29.jpg)
Data are a key resource in
the digital economy.
However, data are not ‘oil’.
Data are human-made
matter (artifacts).

5. A Stream of ‘Artifactual’ Data Studies
•Data sourcing (Jarvenpaa and Markus 2020), repurposing (Aaltonen and Tempini 2014), transforming or
‘warping’ (Stelmaszak et al. forthcoming), and data maintenance and governance for different purposes
(Jarvenpaa and Essén 2023, Parmiggiani et al. 2022)
•Knowledge claims and performativity in data production (Aaltonen and Stelmaszak 2023, Lebovitz et al.
2021)
•The use of specific types of data (Østerlund et al. 2020) and the creation of more complex data-based
objects (Alaimo and Kallinikos 2020), or commodities out of rudimentary data (Aaltonen et al. 2021)
•What makes data ‘data’ for certain knowledge claims (Monteiro and Parmiggiani 2019, Østerlie and
Monteiro 2020)
•How data-based resources are involved in reconfiguring organizational knowledge and knowing (Alaimo and
Kallinikos 2022, forthcoming, Monteiro 2022)
•Data as semiotic material (Bailey et al. 2012, Mingers and Willcocks 2014)

6. Recent studies examine the production, sourcing, use, and
reuse of data in practice.
It is equally important to understand how and why certain
events, entities, and their attributes are recorded (or not) as
data the way they are–we call this data design.

7. ActorToolsRepresentation
Data
Data-
producing
arrangement
Data-producing
arrangement emerges from
local negotiations on how
to capture observations about reality.

8. 1.Data design encompasses different activities involved in forging the way in
which reality is encoded in digital signs. This always entails practical decisions
and compromises.
2.The study of data design must attend both to the computational and semantic
aspects of data.
3.The design of a new or improved type of data to capture organizationally
relevant events, entities, and their attributes can amount to data innovation.
4.Data design can be considered innovation if it uniquely adds value to business,
for instance, when a new or improved type of data enhances the capacity of an
organization to innovate or intervene in its internal and external environments
(Aaltonen and Penttinen 2021).

9. Data Innovation Lens
Studying data design as innovation

10. Algorithms do not live in the
real world.
The better data we have, the
better we can intervene in the
real world with algorithmic
action.

11. The Value of Data Innovation
Economic efficiencies as novel data make new kinds of analytics and
operations possible (Asatiani et al. 2021)
Strategic advantage as novel data lay ground for new products and
services (Aaltonen et al. 2021, Kallinikos and Tempini 2014, Park and
Gil-Garcia 2022) and business models (Wiener et al. 2020)
Improved social justice as we illustrate though a case vignette

12. How to Study Data Innovation?
1.Focus on the creation and alteration of grammatic rules for data
production in practice.
2.Focus on how data evolves over time: data tend to express strong
path dependencies–data innovation can be studied as path creation
episodes.
3.The special role of interface data that acts as the medium that bring
different components of a modular system together.

13. …but Bowker and Star!
Research on classification systems and standards is important reference
point to the study of data innovation. However,
1.Actual data-producing arrangements mix multiple classifications whose
adoption and adaptation is always a local choice.
2.Classification studies are pessimistic about the possibility of improved
classifications: “every standard and each category valorizes some point of view
and silences another” (Bowker and Star 2000, p. 156), whereas data innovation
lens assumes that it is often possible to produce better (if not perfect) data.
3.A classification can emerge bottom-up, from a specific data production effort.

14. Case Vignette
Social justice through data innovation

15. The Gender Identity Data Project
The GI Data Project at a large U.S. public university aimed to allow
students, staff, and faculty to declare their gender identities in data
beyond the traditional male or female binary.
•Technically simple extension to a system that allows campus community
members to declare their identities.
•The result is not the ‘right’ but a better way to record the gender identity of
community members.
•The changes had repercussion throughout the university systems and allowed the
institution to be part of shaping the data collection for other institutions too.

16. Further Reflections on the Case Vignette
1.The anti-modularity of interface data. The role of data as the medium that
defines system modularity is not well understood.
2.Data are not just digital artifacts (Faulkner and Runde 2019, Kallinikos et al.
2013, Yoo 2012). There is a need to attend to both their computational and
sematic aspects, yet our current research traditions may not be perfectly
aligned with such a task.
3.Data innovation can be about breaking free from a taken-for-granted
classification system.

17. Research Agenda
Toward contributions to established streams of research and
deepening the understanding of data innovation

18. Contributions to Established Streams of Research
Research focusLiteratureIllustrative research questions
ModularityDigital innovationHow does data innovation shape the evolution of modular layered architecture?
How do organizations deal with the anti-modularity of interface data?
How do redesigning data flows between modular components contribute to other types
of innovation?
Data qualityData managementHow do data innovation and data quality affect each other?
How does data quality evolve in organizations and beyond over time?
How does the connective quality of data relate to more traditional measures of data
quality?
Data justice Social justiceHow can data innovation reduce potential data injustice?
How does the definition of just data evolve over time?
How can attempts to innovate data further social justice goals?

19. Understanding Data Innovation
Research focusIllustrative research questions
Data designHow are data designed and updated in organizations in practice?
How does data innovation drive further innovation?
How do different types of rules interact in data innovation?
Data (re)combinationHow does data innovation enable the creation of more complex objects of organizational knowing and acting?
What skills and organizational capabilities are required to enable data innovation?
Data evolution What kind of factors support or constraint data design and, hence, innovation over time?
Why some organizations are successful in data innovation whereas others are not?

20. Thank You!
For more on the study of data, see https://datastudiesbibliography.org

21. The Data Studies Bibliography is a curated, searchable bibliography of
papers that focus on data as an object of research. The bibliography is
available at https://DataStudiesBibliography.org and maintained by
Aleksi Aaltonen (Temple University) and Marta Stelmaszak (Portland
State University).
Just added!

22. But Hasn’t Data Design Been Studied?
The ways in which data-producing arrangements account for data have been described as
‘encoding’, that is, as a process by which a system captures data about a phenomenon
along attributes defined by system designers (Alaimo and Kallinikos 2017), as ‘structuring
capacity’ that embeds individual datums into a meaningful data structure (Aaltonen and
Penttinen 2021), and, already while ago, as ‘grammar’ by the rule-based school of data
modeling (Hirschheim et al. 1995).
However, none of these nor the venerable conceptual modeling tradition that develops
tools for data design (Burton-Jones et al. 2017, Recker et al. 2021, Wand and Weber 2002,
2017) or the earlier research on information requirements determination (Appan and
Browne 2012, Browne and Ramesh 2002) pay particular attention to how data are
designed in practice.

23. Are You Reifying Data?
The data innovation lens does not deny the importance of how data are used and
become entangled with organizational practices. Yet, if the data themselves would
not matter beyond how they are put in practice, there would hardly be a reason to
collect the data in the first place.
The data innovation lens extends a broadly practice-based perspective to
understanding to the creation of data-producing arrangements.

24. Presentation history
DateInstitution / eventTitle
May 10, 2024The London School of Economics and
Political Science
Data Innovation Lens: A new Way to Approach Data Design as Value Creation
May 7, 2024KIN Center for Digital InnovationData Innovation Lens: A new Way to Approach Data Design as Value Creation
April 30, 2024ESSEC Business SchoolData Innovation Lens: A new Way to Approach Data Design as Value Creation
May 12, 2022Aalto UniversityBeyond the Facts: Data as Digital-Semantic Artifacts
March 15, 2022University of Jyväskylä (CPSS Lunch Seminar)Beyond the Facts: Data as Digital-Semantic Artifacts
June 8, 2021University of Oslo (research seminar)Data and Value: Digital Data as Technological Artifact