Data management risks

IoT is one of the most significant sources of big data among social media sites, sensor networks and machine-to-machine solutions. Sivarajah, Kamal, Irani & Weerakkody (2017) present in their research three main challenges regarding big data that are data challenges, process challenges and management challenges. Data challenges are caused by the characteristics of data. Process challenges are related to handling the data and preparing it for usage. Management challenges deal with issues regarding for example the governance and the safety of IoT. (Sivarajah et al. 2017, 263, 265) More information regarding these challenges is provided in the figure 9.

Figure 9. Data management challenges (adopted from Sivarajah et al. 2017, 265)

The characteristics of big data make it successful, but they also create some challenges around it. In this thesis 7 V’s of data are used to describe the big data. The amount of data collected through IoT devices is enormous which creates different kinds of challenges. Also, the variety of data sources leads to a situation where the collected data is heterogenous and the possible interoperability issues caused by them can affect the adaptation of IoT and the integration of new systems to existing processes. The data may lack binding information that complicates the data processing and usage. Thus, the expected benefits can be significantly smaller than planned. In addition, the heterogeneity of data can lead to, for example, data quality issues and therefore, misleading information. Veracity issues refers to the fact that both, structured and unstructured data, can have reliability issues. Velocity is about the fast rate of data inflow with non-homogenous structure. Variability of data, referring to the constantly and rapidly changing meaning of data, creates big challenges for example, the analyzing processes. With the help of visualizations, the collected data can become easily readable and therefore, make the data related processes more efficient.

However, the visualization of data can be difficult due to the large sizes of the data sets.

Management

Finally, extracting knowledge from data can create value for companies. This makes data such an important asset for them. (Sivarajah et al. 2017, 269, 273-274; Brous et al. 2020, 5, 13)

The collected data is often so raw that it is hard to understand as such. In addition, the data can be different in terms of subjectivity and importance, and it can be anything from individual opinions to specific measurements. Examining, understanding and finding data patterns is crucial when huge volumes of data from several different sources are collected.

The key challenge of IoT solutions is the intelligent integration of these large data sets in order to create new knowledge. Combining data from several data sources is beneficial as the same information may not emerge from analyzing separately the data sources. The data needs to be pre-processed through data cleansing. Having data from multiple sources can also cause noise and uncertainty which needs to be handled properly. The main objective of these processes is to help companies and users to utilize multi-source data to find useful information through mining and analysis. Structuring data in a way that it is easily and efficiently available for queries and different analysis tools is very important. (Karacapilidis, Tzagarakis & Christodoulou 2013, 227-228)

Based on the Sivarajah et al. (2017) research the process challenges are data acquisition and warehousing, data mining and cleansing, data aggregation and integration, data analysis and modelling and data interpretation. There challenges are mainly caused of the 7V’s of big data. Data acquisition and warehousing problems are mainly caused by diverse data sources and the volume of data. Data mining, cleaning and analyzing can be challenging especially because of the noisy, dynamic and heterogenous nature of the data.

The growing amount of unstructured data sets can cause integration challenges for companies. (Sivarajah et al. 2017, 273-274; Chen, Chen, Du, Li, Lu, Zhao & Zhou 2013, 161)

The sensitivity of data requires careful management. Besides privacy and security challenges there are other management challenges as well. The usage of data continues to increase which leads to growing need of data governance. Data governance means assuring data quality and ensuring that the value of data remains as a company asset. At the same time, the volume of data increases, leading to higher demand for data centers and thus, the workload in them is going to increase. The growing amount of data can lead to, for instance, security, privacy, management and server technology challenges. Data storage and processing costs can be very high. Data ownership challenges are complex, and they need to be solved in order to unleash the full potential of data. Data ownership

issues are often discussed, for example, in social media context as users create the content but basically the social media platforms can have rights to it. Information sharing is beneficial for service supply chains. Companies usually have their own data warehouses based on different kinds of platforms and technologies. The diversity of these systems and privacy regulations may lead to a situation where the actors are reluctant to share their data.

(Sivarajah et al. 2017, 274-275; Otto 2011, 61; Lee & Lee 2015, 436)

4 METHODOLOGY

Before the empirical part of this research, in this chapter the used methodology is presented.

First, the research context and more information about the testbed project will be presented.

This study is conducted as a qualitative research and therefore, the method will be shortly explained. In addition, the data collection and analytics processes are introduced. The interviewees and case companies are going to remain anonymous in order to guarantee openness in the discussions. However, short case company and interviewee presentations will be given in order to provide some background knowledge about them. Also, the secondary data sources will be described. Finally, the process of evaluating the research quality will be examined.