SUBJECT : Big Data Analytics (21CS71)
Assignment 1
USN : 4JN21CS187 NAME : VISHAL S NAVALE
Paper 1 : Big data analytics and the use of artificial intelligence in the services industry: a
metaanalysis
This study adopts a Meta-Analytic Structural Equation Modeling (MASEM) approach to
explore the effects of Big Data Analytics (BDA) and Artificial Intelligence (AI) on service
performance across various industries and regions. The paper focuses on synthesizing
findings from prior research to address three central questions: the influence of environmental
and organizational factors on BDA adoption, the effect sizes linking BDA to service
performance, and the moderating role of AI in enhancing BDA outcomes. By employing a
meta-analytic framework, the authors aim to identify patterns, gaps, and contextual nuances
in the adoption and effectiveness of BDA and AI technologies in the service sector.
The data collection process involved an extensive review of academic journals and peer-
reviewed studies. The authors focused on selecting studies that specifically examined BDA,
AI, and their impact on service performance. Only articles from high-impact journals were
included to ensure quality and relevance. The collected studies spanned a variety of countries,
including both Eastern and Western regions, and covered diverse service industries such as
retail, IT, healthcare, and hospitality. This broad sampling allowed for meaningful
comparisons across different geographic and sectoral contexts.
Once the studies were gathered, the authors employed a coding process to categorize key
variables. Theoretical frameworks used in the primary studies, such as the Resource-Based
View (RBV), Dynamic Capabilities, and Technology-Organization-Environment (TOE)
Framework, were systematically coded. The analysis also identified relevant moderators,
including environmental dynamism, resources and capabilities, and competitive pressures. A
distinction was made between studies focusing on BDA alone and those incorporating AI,
referred to as BDA-AI.
The core of the analysis involved applying meta-analytic techniques to aggregate effect sizes
from the selected studies. The MASEM approach allowed the authors to test relationships
between variables and assess the moderating effects of contextual factors. The study
particularly focused on understanding the role of geographic location and journal ranking in
explaining variations in effect sizes. Notably, the analysis revealed that studies from Eastern
countries reported stronger effects of BDA adoption on service performance compared to
their Western counterparts.
A critical part of the methodology involved addressing publication bias, a common issue in
meta-analyses. The authors conducted a sensitivity analysis to account for potential biases
introduced by high-impact journals favoring studies with significant findings. This step was
essential to ensure the robustness of the results and to highlight the potential overestimation
of effects in top-tier publications. The findings underscore the need for caution when
interpreting results from such journals and emphasize the importance of considering studies
with smaller, yet meaningful, effect sizes.
Finally, the authors conducted a comparative analysis of studies across different regions and
industries. This analysis highlighted the importance of contextual factors, such as digital
infrastructure, market maturity, and organizational culture, in shaping the effectiveness of
BDA and AI initiatives. The study also emphasized the need for managers to tailor their
strategies to local conditions, recognizing the unique challenges and opportunities present in
different markets.
Key Findings and Conclusions
The study concludes that the integration of AI into BDA, referred to as BDA-AI, significantly
enhances service performance compared to BDA alone. The findings reveal that the
combined use of BDA and AI allows organizations to better leverage data for decision-
making, operational efficiency, and customer engagement. The study highlights that Eastern
countries tend to benefit more from BDA adoption, likely due to their focus on digital
transformation and innovation-driven growth strategies. This regional disparity underscores
the importance of contextualizing BDA and AI strategies to local market conditions.
Moderating factors such as environmental dynamism, resources, and competitive pressures
play a critical role in determining the success of BDA and AI initiatives. The study found that
firms operating in dynamic environments are more likely to adopt BDA and AI, as these
technologies help them respond to rapid changes in market conditions and customer
preferences. Additionally, the availability of technical and managerial skills within an
organization is crucial for the effective implementation of these technologies.
From a practical standpoint, the study offers several key insights for managers. It emphasizes
the importance of integrating BDA and AI into strategic planning processes to enhance
decision-making and improve service performance. Managers are encouraged to invest in
technical skills, infrastructure, and knowledge-sharing initiatives to fully realize the potential
of BDA and AI. The study also highlights the need for a localized approach to technology
adoption, taking into account regional differences in culture, economic conditions, and
technological readiness.
The paper also draws attention to the issue of publication bias and its implications for
interpreting research findings. It notes that top journals often prioritize studies with stronger
effects, which can skew the understanding of the true impact of BDA and AI. The authors call
for a more balanced approach to publishing and emphasize the importance of considering
studies with smaller effect sizes, which may offer valuable insights despite their modest
statistical significance.
In conclusion, the paper underscores the transformative potential of BDA and AI in the
service sector, while highlighting the importance of contextual factors and methodological
rigor. By synthesizing findings from prior research and offering practical insights, the study
provides a valuable roadmap for managers and scholars seeking to harness the power of BDA
and AI to drive service performance.
Paper 2 : Impact of Big Data Analytics on People’s Health: Overview of Systematic
Reviews and Recommendations for Future Studies
This paper presents a systematic review of the use of big data analytics and machine learning
(ML) in healthcare, focusing on their impact on public health and clinical outcomes. The
authors employed a comprehensive search strategy, covering multiple databases to include
relevant studies published up to 2020. They used strict inclusion and exclusion criteria,
assessing the quality of each study and extracting data in a structured manner. The review
primarily examines the role of big data tools, such as ML algorithms and decision support
systems, in addressing noncommunicable diseases (NCDs) like cardiovascular diseases,
diabetes, cancer, and mental health conditions. Despite the growing body of research, the
authors find that studies linking big data applications to actual clinical outcomes remain
scarce, emphasizing the need for more evidence in this area.
The authors conducted a narrative synthesis of the studies, evaluating the advantages,
challenges, and future opportunities of big data in healthcare. They found that, while big data
tools hold promise for improving disease prevention, diagnosis, and management, the
evidence on their impact in real-world healthcare settings is still limited. A significant
challenge identified is the variability in research methodologies, with many studies lacking
standardized protocols for reporting results. The authors call for standardized methodologies
and more rigorous reporting to improve the reliability and reproducibility of research in this
field.
The authors stress the importance of developing data standards and improving collaboration
between healthcare providers, researchers, and policymakers to address these issues.
Furthermore, they emphasize the need for transparency in reporting the details of ML models,
including data preprocessing, feature selection, and validation methods, to ensure that the
results can be replicated and trusted in clinical settings.
In conclusion, the paper provides a detailed overview of the current state of big data and ML
applications in healthcare, underlining the significant potential of these technologies while
acknowledging the gaps in research and practical challenges. The authors call for more high-
quality studies that focus on the real-world impact of these tools on clinical outcomes,
improved standardization in reporting, and better integration of big data solutions into
healthcare systems. With more focused efforts, they suggest, big data and ML could
significantly enhance healthcare delivery and patient outcomes in the future.
SUBJECT : Big Data Analytics (21CS71)
Assignment 1
USN : 4JN22CS412 NAME : PRAMATH S
Paper 1 : Impact of Big Data Analytics on People’s Health: Overview of Systematic Reviews and
Recommendations for Future Studies
This paper conducts a systematic review of the applications of big data analytics and machine
learning (ML) in healthcare, with a particular focus on their impact on public health and
clinical outcomes. The authors employ a rigorous search strategy, which was designed and
executed by an experienced librarian, Maria Björklund, from Lund University. The search
covers multiple databases to ensure comprehensive retrieval of relevant studies published up
to 2020. The study follows a predefined protocol, with strict inclusion and exclusion criteria.
Studies included in the review were assessed for quality, and data extraction was performed
in a structured manner, involving two stages of screening and validation. The extracted data
was then analyzed for patterns and trends related to the application of big data and ML in
improving health outcomes, particularly focusing on noncommunicable diseases (NCDs)
such as cardiovascular diseases, diabetes, cancer, and mental health conditions. By including
a broad range of studies, the authors aim to provide an inclusive and holistic view of the
current state of big data and ML technologies in healthcare.
In terms of research design, the systematic review utilized a narrative synthesis method to
summarize the findings of various studies. This method was chosen to ensure that the
qualitative and quantitative aspects of the included studies could be analyzed in a
comprehensive manner. The focus was to evaluate how big data tools—such as ML
algorithms and decision support systems (DSS)—are being implemented in healthcare
settings. The authors looked into the advantages, challenges, and future opportunities
presented by big data in public health, particularly regarding disease prevention, diagnosis,
and management. Through this review, they sought to identify whether the integration of
these technologies was leading to improved patient outcomes or enhancing the overall quality
of care. Key studies included in the review were evaluated for their impact on clinical
outcomes, although the authors note that such studies are still scarce. This gap in research
highlights a key limitation in the current body of literature and emphasizes the need for
further exploration of how big data applications can directly influence healthcare delivery in
real-world settings.
The quality of evidence across the reviewed articles ranged from low to moderate, which the
authors attribute to methodological inconsistencies and a lack of standardization in reporting
results. Many studies failed to adhere to a consistent set of criteria for evaluating the
effectiveness of big data solutions, which made it difficult to draw definitive conclusions
about the relative benefits and risks of these technologies. Moreover, the heterogeneity of the
methods used in the studies contributed to the overall variability in findings. This variability
is particularly concerning when evaluating healthcare outcomes, where consistency and
reliability are crucial for ensuring that technologies are not only effective but also safe for
widespread adoption. To address these challenges, the authors stress the importance of
establishing standardized protocols for reporting the results of big data and ML studies,
ensuring that future research can be compared and aggregated more effectively.
Furthermore, the authors point out that many of the studies included in the review did not
provide sufficient details on critical elements such as data preprocessing, feature selection, or
hyperparameter tuning. These elements are crucial for understanding the actual performance
of ML models in healthcare applications. The lack of transparency in these areas undermines
the reproducibility of the studies, which is a key concern for scientific progress in the field of
healthcare technology. As such, the authors advocate for a more rigorous approach to
reporting experimental protocols, with a particular emphasis on providing clear
documentation of the algorithms used, the data sets analyzed, and the validation methods
applied. The inclusion of this level of detail is necessary not only to improve the replicability
of studies but also to build trust in the application of ML and big data technologies within the
healthcare industry.
Additionally, the paper addresses the ethical and practical challenges associated with the use
of big data and ML in healthcare. One of the key challenges is the lack of uniformity in the
data sources used across different studies. Healthcare data is often fragmented, coming from
various sources such as electronic health records (EHRs), wearables, and patient surveys,
which creates issues of data integration and interoperability. This fragmentation complicates
the application of ML models, as the models require high-quality, well-structured data to
make accurate predictions. The authors highlight the need for improved data standards and
better collaboration between healthcare providers, researchers, and policymakers to ensure
that data can be shared and used effectively. They also discuss concerns related to data
privacy and security, particularly when handling sensitive patient information. Given the
increasing reliance on digital health tools, these concerns must be addressed to ensure the
safe and ethical use of big data and ML in healthcare.
Key Findings and Conclusions
The key focus of the paper lies in assessing the current state of big data and ML applications
in healthcare and public health, particularly in relation to the prioritized World Health
Organization (WHO) indicators for noncommunicable diseases and the ongoing COVID-19
pandemic. One significant contribution of the review is its identification of the gaps in
research, particularly the scarcity of studies linking big data analytics to actual clinical
outcomes. Although there is a growing body of research on big data applications in areas like
disease prediction, management, and monitoring, the authors highlight that much of the
existing evidence lacks direct impact on clinical practice. The paper also outlines the current
challenges faced by healthcare systems in adopting these technologies, such as data
fragmentation, security concerns, and issues with data standardization and interoperability.
In conclusion, the paper provides a comprehensive overview of the current state of big data
and machine learning applications in healthcare, with a focus on public health and clinical
outcomes. The authors note that while significant advancements have been made in terms of
the capabilities of big data tools, the evidence for their direct impact on clinical outcomes
remains limited. The low to moderate quality of the evidence, along with the variability in
results across different studies, underscores the need for further research in this field. The
paper calls for more standardized methodologies, better reporting of experimental procedures,
and a focus on evaluating the real-world impact of big data and ML technologies on
healthcare delivery.
Paper 2 : Big data analytics and the use of artificial intelligence in the services industry:
a metaanalysis
This study employs a Meta-Analytic Structural Equation Modeling (MASEM) approach to
explore the effects of Big Data Analytics (BDA) and Artificial Intelligence (AI) on service
performance across different industries and regions. The authors conduct a comprehensive
review of high-impact academic studies that focus on BDA, AI, and their impact on service
performance. The research includes studies from diverse geographic regions, including both
Eastern and Western countries, and various service industries such as retail, healthcare, IT,
and hospitality. The analysis categorizes key variables and moderators, such as environmental
factors, organizational resources, and competitive pressures, to assess how these contextual
factors influence the adoption and effectiveness of BDA and AI.
The findings reveal that the combination of BDA and AI (BDA-AI) significantly enhances
service performance compared to BDA alone. The study highlights regional differences,
showing that Eastern countries tend to experience stronger effects of BDA adoption, likely
due to their emphasis on digital transformation and innovation-driven growth. The paper also
underscores the critical role of moderating factors, such as environmental dynamism and
organizational capabilities, in determining the success of BDA and AI initiatives.
Organizations in dynamic markets are more likely to adopt these technologies to adapt to
rapid changes, and the availability of technical skills is vital for successful implementation.
In conclusion, the study provides valuable insights for managers looking to leverage BDA
and AI for improved service performance. It stresses the importance of integrating these
technologies into strategic planning and tailoring adoption strategies to local market
conditions. The paper also addresses the issue of publication bias, noting that studies with
more significant findings are often prioritized in top journals, potentially skewing the
understanding of BDA and AI's true impact. Overall, the study emphasizes the transformative
potential of BDA and AI in the service sector, while highlighting the need for a
contextualized and methodologically rigorous approach to their adoption and implementation.
