In studying enabling factors for data democratization process, this paper argues that learning from Actor-Network Theory (ANT) teachings can be extended in framing a sustainable data democratization model. According to Gherardi [7] , the ANT forms the umbrella approach known as Practice Theory or Practice-Based Research (PBR). ANT is a theory and a method. PBR approaches are holistic and qualitative practices formed by a set of activities that acquire sense and make them a unit. According to Farias et al., ANT is a way of doing and engaging in the world [8] . Callon et al. [9] also talked about ANT and its contributions to the understanding of the configurations of human and non-human elements that enact organizations and realities. From ANT practice perspective, knowledge, technology, and abstract concepts can be usefully interrogated by examining the human-technology relationships that produce them [10] . The enabling factors are the following four key concepts of ANT: spokespersonship, matters of fact/matters of concern, obligatory passage points and hybrid forums. The core theoretical perspective adopted in this research fits the ANT practice that seeks to retrieve the complexity and heterogeneity that constitute reality as enacted in Law [11] , suggesting that it occurs in a particular prescriptive, and non-problematic way [12] . For ANT, all things are understood as enactments. In the next section, we define data democratization and propose a framework for visualizing data democratization in the context of study design.

Because the main aim of this research is to advance, refine and expand a body of knowledge on data democratization in Cameroon, establish facts, and construct an explanatory framework, the design science approach of Vaishnavi and Kuechler [13] was found most suitable. Our awareness of the problem and initial proposal is based on a qualitative meta-synthesis of data democratization projects in research and practice to reconcile findings across studies unearthed with a structured literature review. Following Beck et al. [14] formula, we extend Vaishnavi and Kuechler [13] procedure with theory-building elements from the interpretative research method grounded theory [15] . Grounded theory (GT) is a structured, yet flexible methodology. The process is iterative and recursive. GT is performed through a systematic data collection procedure, identification of categories (themes), linking these categories, and forming theories that explain the process. According to Chun Tie et al. [15] , theory is not discovered; rather, theory is constructed by the researcher who views the world through their own lens. Because limited research has been carried out on data democratization in Cameroon, this methodology is appropriate. The methodology enables to produce or construct an explanatory theory that uncovers a process inherent to the substantive area of inquiry. The adapted approach consists of four distinct phases: Awareness, data collection and suggestion, development, and evaluation and conclusion. See Figure 1 for a summary.

Although the phases are sequential, they were iterated until a coherent framework emerged. Specifically, we have performed three design iterations of data collection and suggestion, development, and evaluation and conclusion: Iteration 1: Structured literature review to enhance theoretical sensitivity, meta-synthesis, initial framework, Iteration 2: Expert interview study, data coding, consolidated framework, demonstration, and expert feedback, Iteration 3: Framework refinement, evaluation workshops, final framework.

• Awareness. The artifact of this research is a framework to facilitate and guide the introduction of data democratization in companies to aid the systematic design, development, and evolution of implementations.

• Data collection and suggestion. To examine the current state-of-the-art and provide our first contribution, we conducted a rapid evidence assessment of the literature on data democratization. Data democratization is an evolving topic and as Tate et al. [16] suggested, conducting an in-depth literature review aids in understanding the current body of knowledge and identifying the research gaps. The systematic literature review model suggested by Watson [17] and Levy and Ellis [18] and Webster were followed. A literature review according to Cronin et al. [19] , must comprise several sources of data to be relevant and effective. This study considered more than two databases (i.e., AIS virtual library, Wiley online library, Taylor & Francis online library, Emerald insight, Springer, Science Direct, IEEE, ACM digital library, and Scopus), industrial initiatives, business case reports, along with grey literature using search terms “data democratiz(s)ation”, “democratiz(s)ation of

data” and “democratiz(s)ed data”. While we were aware of the limited scientific research in data democratization, we still considered a literature review to structure the domain as important. From the analysis of the articles, we designed a first iteration of the framework. In the second iteration, we conducted semi-structured interviews to verify the first iteration of the framework and adapt it according to the input we received from the interviewees. The interviews were recorded, anonymized, and transcribed. To extract information from the interviews, we coded the transcripts iteratively (using open and axial coding) and analyzed them with the grounded theory approach. Some of the publications were included at a later stage when we conducted a “snowball” search by following up references cited in literature that we included at the initial stage. There was no restriction placed on the discipline on which the article’s focus is on; as data democratization is perceived as a multidisciplinary field. However, only papers written in the English language were considered for this study.

• Development. Based on the evaluation of the structured literature analysis through the interview study, we combined the identified stages and phases from both analyses. The first version of the framework emerged from the results of the literature analysis, which were adapted and supplemented by the expert interviews. The abduction logic casual mapping determination suggested by Nandi et al. [20] and Narayanan and Armstrong’s [21] was followed.

• Evaluation and conclusion. Our evaluation was informed by Venable’s FEDS Framework [22] to “demonstrate the utility, quality, and efficacy” [23] of our design artifact. Due to the nascent nature of our research with the absence of general recommendations for data democratization implementation, we decided to implement a two-staged naturalistic, summative evaluation based on a human risk and effectiveness strategy [22] . First, we presented the framework to the interviewed experts again and collected their feedback. Second, we evaluated the applicability of the revised and refined version of our framework in multiple workshops using real-life cases. We conducted online meetings with multiple companies and applied the framework to their respective situation. We used their input to finalize our framework as outlined above. After the final consensus between the team proposed enablers and subject-matter expert’s feedbacks was reached, an outside case study found in the literature was used to partially validate the framework. The learning from the case study workshops constitutes our third contribution.

This study builds on other recent work in the data democratization discipline. Literature review shows that democratizing data is a multi-faceted, complex phenomenon with many sub-concepts, and definition is thus in flux. In this study, we did not select one definition of democratization as a normative baseline to critically assess our material but strived to induce various definitions of “democratization” from different authors’ writings. As the selected definitions in Table 1 shows, the terms data democratization is used in different ways, and it is agnostic practices.

Different definitions imply the need for interoperability, ease of integration, openness and inclusiveness, trust between those data producers and the data consumers. Establishing a unified definition of the data democratization is the not the goal of this paper but creating a flexible and robust architecture for data democratization is. By combining the various insights acquired through analysis of the various definitions, this study defines the agnostic term “data democratization” as: “a holistic attitude of willing organizations focused on curiosity, learning, and experimentation for delivering trusted data for trusted insights to a broad range of authorized data stakeholders”.

In this section, we analyze prominent scientific research papers, widely used data democratization success factors, and best practices to study the mechanisms, and factors that can be used to gauge and benchmark an organization’s data democratization architecture. Several applications of data democratisation success factors can be found in many different fields such as domain of healthcare (Eichler et al. [38] ; Lewis et al. [39] ; Kuiler & McNeely [40] ; Minielly et al. [41] ), energy resource (Yoder [42] ; DiChristopher [43] ; Husseini [44] ), education (Fay [45] ), housing market (McLaughlin & Young [46] ; Grey [47] ), agriculture (Chandra et al. [48] ). The many recent publication cited shows that data democratization in both theory and practice is gaining ground and acceptance. While data democratization is recognized as an important concept in research and practice, it is still unclear what it comprises and how it is built (Labadie et al.

[49] ; Lefebvre et al. [50] ). Therefore, identifying existing data democratization guide and success factors for data democratization is considered an important part of this study. Much of the cited literature discussed data democratization success factors to be multifaceted. A factor could be an enabler, an obstacle, or both. For example, one of the factors was data leadership. This was presented in the literature as an obstacle if it was lacking, and an enabler if it was present. Table 2 summarizes the most important data democratization success factors or enablers empirical results.

A principal tenet in this HCDDF is the common-sense idea that, at any given time data democratization implementation attitude is an IT capability. It shows questions to be asked and issues to be taken care of. Consistent with the reasoning in our analysis of previous literature and lived experience, critical success indicators were identified using manual coding and grouped into 6 explainable data democratization building indicators: 1) “Organization structure, ethical guidelines, business context and stakeholders’ value”, 2) “Data leadership and data culture”, 3) “Data literacy, continuous training and capacity building”, 4) “Data Observability, trustworthy, standardized data sets, interoperability”, 5) “Sustainable data technologies |platforms, access to data and tools”, 6) “Intelligent Data Governance and Privacy”. These six barrier-enabler couples describe the general prerequisites of sustainable, flexible, adaptable, and practical framework for data democratization solutions.

Figure 2 shows a representation of the data democratization framework, outlining how the different components relate to each other. The framework is

aggregation of indicators describing the different aspects of a robust data democratization attitude and capability. A consideration of all these factors in the development of human-centric data democratization approaches can lead to automated functions. Data democratization capabilities are critical resources for any organization that sees business data as an organizational asset.

Although each perspective prompts different types of research question, they should be thought of as overlapping rather than mutually exclusive. Data democratization must start with human-centric design and continue with human-centric workflows. The initial validation efforts (interviews and focus groups) of the model were a valuable way to capture the perspectives of SMEs who have worked in this field.

Data will talk to you only if you are willing to listen. But not all enterprises are very well positioned to leverage value from their data assets. In the following section, we first present the evaluation results in the context of the data management workshop conducted at B’SSADI GALLERIES. External triggers such as the recent COVID-19 pandemic and the ongoing Anglophone Crisis has exerted multiple pressures on the organization to digitize their business data. B’SSADI GALLERIES has embarked on a journey towards sweeping digital transformation with intention of providing modern data analytic platforms for the ’IT Team’. B’SSADI GALLERIES management indicated their great desire towards becoming data-driven organization and was interested to understand how best to structure their data analytic project. The researcher spoke about how a holistic approach would help in creating a more solid objective and key results system for different departments and not just the “IT Team”. The B’SSADI GALLERIES workshop consists of two primary activities—presentations and three focus groups’ discussion. In the first part of the workshop (two full working days), we presented the data democratization holistic approach framework. The second component of the workshop engages participants with discussions around the state of adoption of data democratization principles in their own institutional contexts and subsequent completion of questionnaire.

The framework (Figure 1) proposes that the data democratization holistic approach maturity of a willing organization must intentionally develop in its six structuring elements in a synchronized way to unfold its data democratization potential. To be able to assess the dimensions of the framework in a standardized and assessor independent way, we formulated hypothesis/questionnaire building on the definitions of each of the data democratization building blocks (see Section 3.3). The questionnaire describes scenarios for each maturity level and each assessed dimension. The questions focus was to discover where B’SSADI GALLERIES’s employees believed that they are in incorporating the various principles of democratizing data to achieving their desired goal of becoming a data-driven organization. Each dimension is loaded with check questions reflecting its constituents. For example, the first dimension has the following required capabilities: “data ethics,” “business context” and “stakeholder value”.

Across the six dimensions, respondents rated B’SSADI GALLERIES data implementation on a scale of 1 (low or absent/ad hoc related to the capability; it is addressed in an improvised, irregular way) to 5 (high or capability performance is regularly assessed to improve practice and manage risks) as defined in Table 3. The test questionnaire poses 20 questions separated into five different categories. That is, we applied the principles in Figure 1 and evaluated these using criteria in Table 3. While assessing a project one can give the project a score from one to five for each question, sum the total score and divide it by 100. By doing this, one is left with a score between zero and one, representing how data democratic the process of data analytics is.

Maturity level 1 (Absent/ad hoc) and 2 (Repeatable) as defined above are mostly where the organization is on their data driven requirements. On the question of if their approach ensured that data usage is optimized to effectively realize business objectives without running afoul of ethical data usage guidelines and regulations. The measured maturity is registered at level 1. We had equally concluded based on the previous focus group discussion that there is a lack of understanding of what data ethics should constitute of among the employees. Employees are not able to independently operate systems where they are available beyond their standard functionalities to satisfy their information needs, e.g., analytical applications required for their daily work. The participants show

from their maturity level selection that they do not have standardized capability and have documented procedures organizing, educating, and creating tools that enable non-experts to become involved in the process of governing with data. Yes, data democratization demands data access to and for everyone, but the process also requires intelligent checks and balances in form of governance mechanism. In addition, an important part of changing systems which data democratization approach advocate is changing attitudes. Over 80% of the participants gave indicated that they are either Maturity level 1 (Absent/ad hoc) or 2 (Repeatable).

Another hypothesis that we tested is that almost all data analytics projects require a collaborative, multidisciplinary team approach that needing training programmes to equip all data stakeholders with the multidisciplinary networks and skills that are required to develop and/or use data technologies. Participants indicated that the current training provision is not fit for purpose; with over 60% of the participants indicating that they are at maturity level 1 (Absent/ad hoc) and 25% saying maturity level 2 (Repeatable).

Another enabling factor the hypothesis question tested is the issue of data culture. Besides formal processes, organizational structure and assigned responsibilities, the culture among the associates has a paramount influence on how the concept of data democratization is embraced in a data-driven decision-making organization. Like most of the other questions, B’SSADI GALLERIES as an organization is operating at maturity level of 2 with most “decisions adapted case by case on the basis of personal knowledge or the intuition of the manager”. It must be emphasised that this is a preliminary analysis, and for formal recommendations, a much fuller and more systematic analysis would be required using the principles set out in Figure 1 and Table 3.

This study used qualitative approach. Qualitative methods allowed us to explore and map the idea of data democratization or democratizing data topic in detail. Our mapping of the characteristics and enabling factors also has limitations, which need to be considered to qualify our results. First, the use of peer-reviewed articles and documents in English limited our attention to comparatively privileged voices in the discourse on data democratization in digital economy. A more thorough mapping of visions of democratization, which could capture alternative visions of democratizing data, would need to extend the materials to other languages and materials. It should be noted that research into human evaluation of the framework is piecemeal and non-systematic. It relies heavily on self-reported behaviors and beliefs which are not always reliable.