Big Data Analytics Research Methodology

Methodology

Discuss about the  Big Data Analytics Research Methodology.

The current society has been surrounded by so many problems, some which can be solved easily, others need a keen analysis and others which cannot be solved. Scholars have come up in the event of trying to look for clear routes of solving the many existing problems which come up on day to day basis. Many research methodologies have been proposed, therefore, some which are still under investigations(Denzin, 2017). However, everything that is introduced must always have pros and cons, so the existing methodologies have.

In this research paper, we have analyzed a sample of existing research methodology and how it can be used to solve the various research problem. We have also selected a proposed methodology which seems to be more efficient and explained how it can be used to solve a sample problem. Lastly, a comparison of the two strategies has been contrasted to identify their strengths and weakness and a conclusion of why the chosen methodology tends to be the most efficient research methodology.

There exists traditional and modern research methodology, however, traditional methodologies cannot be used to solve big data analytics(Bryman, 2016, p.4). Therefore, the purpose of this paper is to identify an appropriate methodology to big data analytics.

A methodology is simply a system of methods used to a particular area of study in solving the existing problems. A sample of well-known methodology is Quantitative and qualitative analysis. Some other methodologies that have been used by scholars in research include case study, archival research, content analysis and event sampling methodologies(Dawn, 2018, p.95). This methods are used for various problems depending on the nature of the problem. In our work, we will concentrate on quantitative and qualitative analysis. 

According to(Chen, and Zhang, 2014), big data is the application of specialized methods and technologies in the processing of large sets of data. These datasets are very cumbersome and cannot be analyzed by traditional methods of analysis. Examples are web blogs, call records military surveillance, video archives and medical records. The analysis of these data has been an existing problem to scholars since their study usually has a lot of information which can cause a lot of confusion and time consuming.

Research problem statement

Computer simulation which is a modern research method is more efficient that quantitative and qualitative analysis which are traditional data analysis methods

Research problem

Literature review

Many scholars have come to conclusion that big data analysis has grown to be a current problem over the past years, traditional methods of data analysis have failed to a big extent to substantiate some of the data analyzed. This has raised a major concern leading to introduction of modern methods which are computer generated. These modern methods only need filling of data in computer then all the tasks are completed. This makes the entire task simple and can be completed only by a mouse click. It also saves time since a lot of data can be analyzed, charts constructed within microseconds. However, these methods requires experts who have done a lot of study in modern technology, information systems and with computer programing skills. As a matter of fact, the computer simulation strategy has also had some disadvantages. It has costed organization a lot of resources just to analyze single sets of data. As compared to traditional data analysis methods which analyze data manually, the modern computer simulation strategies frees organization with the number of employees who have to be employed to analyze the specific data in quantitative and qualitative methods, the data can be analyzed as shown below.

Research Problem	Sub-problem	Collected Literature
Problem 1	Sub-problem 1	Literature 1
Analysing big data	Traditional methods cannot effectively complete big data analysis due to their complexities	In the article management revolution by (McAfee, Brynjolfsson, Davenport, Patil, and Barton, 2012). Traditional methods are cheap and effective but cannot analyse big data while on the other hand modern methods have also few challenges since they are handled by computer generated algorithms.
Problem 2	Sub-problem 2	Literature 2
Difficulty in translating big analysed data	Big data need modern computer simulation methods which are expensive to implement	(Hashem et al, 2015) in the rise on big data in computing states that modern methods of data analysis can effectively complete big data within seconds but are expensive in nature.

Quantitative approach methodology is a type of strategy which surveys, structures and observes records of numeric data, it is based on a deductive process which tests concepts or the problem, it is more objective and number based(Dawn, 2018, p.95). The method has fixed response options with statistical tests being used in analysis and largely depends on measurement device used. The qualitative method focuses on groups, for instance, interviews, it is inductive in nature and is text based. This strategy has an unstructured response and has no statistical tests. It therefore largely depends on skill and rigor of the researcher. This two methods are traditional methods and cannot effectively analyze bid data(Dawn, 2018, p.98).

Qualitative methodologies: Examples of qualitative research types include Historical which establishes facts and draws conclusions concerning past events, comparative methodology which are used with historical research to compare peoples experiences, action research type that is similar to experimental research which is conducted in the real world and lastly case study research methods entailing an empirical inquiry which investigates a contemporary problem in-depth 

Quantitative methodologies: On the other hand, quantitative methods include correlation research types which describe statistical measures association and relationships between different phenomena, experimental type, this isolates, and controls relevant conditions which determine the investigated event. Lastly, in this category we have evaluation research approach, this makes judgments concerning the merit or worth of educational programs or products.

Types of Methodologies

Since big data needs modern methods of data analysis, Computer simulation can effectively serve this purpose. This is because the researcher despite having a lot of information, only needs to fill in the details and the computers arrange, analyses, interprets and draws tables and charts for the submitted data.

Computer simulation is system based algorithms used to compute data by producing the behavior of systems using mathematical models. They can run both on small and large scale basis depending on the data provided(Hertz, 2018, p.54). These systems can analyze millions of information by just having a mouse clicked, drawing data tables and interpreting the data for the researcher. The types of models used to compute data include stochastic steady state, continuous or desecrate and dynamic system simulation. The computer-based algorithms in this modern method can explore massive input spaces, uncover interesting features of complex variables in response surfaces and explicitly identify cause-and-effect of data relationships. 

Sample Problem

The population census is conducted in many countries, Australia is one of them. The data generated from counting is always cumbersome and cannot be analyzed manually.

How Selected Problem can Solve sample problem

After counting the number of people, the data can be analyzed by computer simulation strategy by just filling in the results. The computer will count total populations in all regions and give figures which can be compared with previous years and analyzed. With this, it is easy to get the trend in population for the current year.

Relevance of selected Methodology to the selected research problem

Computer simulation does not only analyses the data but also interprets them, draws graphs, compares data and most importantly keeps the information for future references, something that the traditional methods cannot do.

Comparison between the two Methodologies

Quantitative and qualitative methods Vs Computer simulation: The main difference between this two approaches is very clear. Computer simulation can do all the task that quantitative and qualitative analysis can do but the later cannot do what computer simulation canthus includes completion of massive data analysis within seconds as well as keeping records for future references (Chambers, 2017, p.34). Additionally, computer simulation is less biased as compared to what the other methods can be.

Proposed Methodology analysis

Computer simulation can effectively conduct massive data research analysis within a few time. This strategy of the data analysis(Hashem et al, 2015). The strategy involves the following generic steps. Research question identification, a model design where the target is specified, model building and model verification. After the model has been verified, the simulation is run after which model validation is used. The following chart shows this basic steps 

Selected Methodology

It is therefore evident that computer simulation approach can be an effective method of huge data analysis basing on the fact that it can analyze, classify, and interpret massive data within a very short time.

Strengths and Weaknesses

This methodology saves time and resources needed to hire human personnel to analyze the data since all analysis is done on a computer. Additionally, the data ones entered in the system is secure and can be preserved for future references. The computer-based analysis also reduces chances of the data being manipulated due to self-interests of researchers. However, the strategy needs qualified personnel who have skills in computer programming and information technology and acquiring such people is also expensive. In the event that the system breaks down due to hackers’ activities, a lot of data is lost(Mandrà, Zhu, Wang, Perdomo and Katzgraber, 2016). Lastly, computers entirely depend on the programmers who are people and have the power to manipulate figures.

Framework

In our sample problem, for instance, the counting of people in Australia, the framework has independent and dependent variables. Staring from our problem which is analyzing the data, the researcher depends on the sources of data which is population sample. He also depends on the methodology which will be used to analyze the data using computers. However, the computer also depends on the researcher since he figures in the data. Additionally, the analysis of the computer and validity depends on the validity of the data sources(Chambers, 2017, p.49). When all these elements play together without manipulations, an effective data analysis takes place. If it does not, the result cannot be changed apart from doing another analysis. That is why we have one independent variable, which cannot be changed as shown below.

Conclusion

Big data can be studied with the help of computer simulation methodology as opposed to other traditional methods of data analysis. The strategy has though some weaknesses which exposes data to threats for instance hackers. Additionally, the efficiency of the analysis in computer simulation entirely depends on the model used in analyzing the data(Chambers, 2017, p.48). Various elements of research affect the validity of data, it is the role of the researcher to control all the dependent variables in order to ensure the result projected are not biased.

This study, however, has only analyzed two methods of data analysis. There are other methods which can as well work better with big data analysis. Additionally, this study does not imply that qualitative and quantitative approaches cannot manage massive data analysis at all. The differences between the two approaches and computer simulation are only efficiency and reference in future. 

References

Bryman, A., 2016. Social research methods. Oxford university press.

Chambers, J.M., 2017. Graphical Methods for Data Analysis: 0. Chapman and Hall/CRC.

Chen, C.P., and Zhang, C.Y., 2014. Data-intensive applications, challenges, techniques, and technologies: A survey on Big Data. Information Sciences, 275, pp.314-347.

Dawn, R., 2018. Quantitative and Qualitative Analysis. In Educational Achievement and Psychosocial Transition in Visually Impaired Adolescents (pp. 95-110). Springer, Singapore.

Denzin, N.K., 2017. The research act: A theoretical introduction to sociological methods. Routledge.

Hashem, I.A.T., Yaqoob, I., Anuar, N.B., Mokhtar, S., Gani, A. and Khan, S.U., 2015. The rise of “big data” on cloud computing: Review and open research issues. Information Systems, 47, pp.98-115. Hertz, J.A., 2018. Introduction to the theory of neural computation. CRC Press.

Mandrà, S., Zhu, Z., Wang, W., Perdomo-Ortiz, A. and Katzgraber, H.G., 2016. Strengths and weaknesses of weak-strong cluster problems: A detailed overview of state-of-the-art classical heuristics versus quantum approaches. Physical Review A, 94(2), p.022337.

McAfee, A., Brynjolfsson, E., Davenport, T.H., Patil, D.J. and Barton, D., 2012. Big data: the management revolution. Harvard business review, 90(10), pp.60-68.