information Study

Date: 12-4-18

EFFECT OF BIG DATA ON NIGERIAN BANKS

Abstract

In the current bank industry, most of the data has now been transformed into what is currently known as big data. Some banks have now started taking advantage of big data in order to satisfy the main objectives entailed in their marketing campaigns. The banking industry can use the data in order to increase their efficiency bythrough the identifyingication of the key customer, improving the customer feedback system, detecting when a customer is about to be lost, enhancinge the active and passive security system, ands well as efficiently evaluating the system. Comment by Author: You need to support generalizations like this.

This paper focuses on the effect of big data on Nigerian banks as well as the different analysis and algorithms that the banking industry can use to achieve all the advantages of big data especially in the Nigerian banking industry. Analysis such as link analysis, survival analysis, neural analysis, text analytics, decision tree, clustering analysis, sentiment analysis, Datameer and social network analysis for the prediction of security threat. (Widrow, Rumelhart and Lehr) Comment by Author: This is not a sentence

Introduction

The main issues faced in the area of banking in emerging markets are oftentimes similar to those experienced elsewhere but these challenges are encountered when fewer resources as well as structures present to properly address them. Key issues experienced by most of the industry include expanding the market as well as improving the financial inclusion in order to provide better access to the banking services. One of the most intriguing aspects of the current market is the growth and emergence of fintech and especially how it is greatly assisting the people who were previously hindered from accessing banking and getting loans. Comment by Author: In a doctoral research paper, you must support generalizations with citations to the literature. Comment by Author: Source? Comment by Author: Source?

“Fintechs” can be used to champion the providers who expand the market. They are not necessarily being used to replace existing banks and services but they are being used to provide new services to customers who had been notoriously underserved for much of the past. This is key in areas such as Nigeria. This is due to the fact that the government’s vision for 2020 is for them to join the G20 group that represents the world’s top economies. The Nigerian government views digital banking and fintech as one the major ways that they can achieve this and they hope that they can transform the country into Africa’s fintech centre in the process. (Widrow, Rumelhart and Lehr)

“Big data” is described as a term that can be used to describes large volume of data: both structured, semi structured and unstructured data. Typical bank data is morphing into big data due to the sheer magnitude it possesses. It has a large volume (terabyte record, transaction, tables and files). Additionally, the velocity is a key factor. The velocity includes batch, real time and stream. Moreover the banks also consider variety. This includes structure, semi-structure, unstructured data as well as all of the above. All of these factors have exceeded the abilities of what an IT system can ingest, store, analyse and process in real time. (Widrow, Rumelhart and Lehr) Comment by Author: Source? More data does not equal a “big data” approach

Big data does not necessarily fit into tables of columns. There are various new data types such as the ones stated above e.g. unstructured and semi-structured. These data types can be processed in order to yield insight into a business or conditions like data obtained from twitter feeds, blogs and even third party data. Additional factors include call detail report (audio format), network data video that is obtained through surveillance cameras and equipment sensors, which are oftentimes not stored in a data warehouse. Comment by Author: Source?

In 2012, Gartner updated the definition of 3V model (This is variety, volume and velocity) to high variety, high velocity and high volume. This change was made due to the tedious nature of data collection and analysis in a big sector such as the bank industry. The gathering of large amounts of data from various sources making big data very alluring for the bank industry. Big data is extremely powerful for the bank industry as it allows for the making of effective decisions faster and better than traditional business tools. The technology behind big data is known as Hadoop (). (Widrow, Rumelhart and Lehr) Comment by Author: ? Comment by Author: Not a sentence Comment by Author: Source

Nigerian banks stand to gain a lot with their implementation of big data in their banking system. It allows for customer segmentation. This is a way in which one can group customers to share the same interest together. Big data also allows for risk management through the prevention of fraud. With big data, no unauthorized transactions can be made and massive information plays a critical role in the desegregation of a bank’s needs into a platform that is both practical and centralized. Furthermore, big data allows for efficient evaluation of the system. Through the proper use of big data, the banking industry will be able to provide an exact evaluating report. (Noori) Comment by Author: At this point, I don’t have a clear idea about what your research topic is? What is your original contribution that makes this paper publishable in a peer-reviewed journal? Comment by Author: Source?

Big data also allows for the improvement of customer feedback. The Nigerian banks will be able to identify the exact problem a customer is dealing with in real time. It can additionally provide the accurate services in a prompt fashion. The banks in Nigeria will also be able to identify the key customers to the bank. Moreover, the banks can enhance their security through the implementation of big data. They can easily detect fraud and also boost the active and passive security. Big data can also allow for detection of customer exits in the banking hall. The banks can get easily notified of this. Additionally, big data allows for sentiment analysis which assists the bank in monitoring customer satisfaction. (Chunsheng Zhu) Comment by Author: Source? I would think you do this before you enter transaction data into a data warehouse Comment by Author: Source?

The big information revolution that is taking place in the 21st century has found a deep resonance with money service corporations in the Western coast of Africa. It is additionally appealing if one considers the dear information the banks have been storing for several decades. (Desai and Kulkarni) Comment by Author: As you’re shifting to the literature review, I still don’t know what your original study is. I need to know this in order to assess whether or not you’ve selected the appropriate literature.

Literature Review

Banks square gauge establishments that operate in the sphere or domain of money/business. This area concerns activities like disposal, management of deposits and capital market investments among others. The banking system in Nigeria is integral for the economy and thus it is a topic of interest/draw from researchers in an exceedingly wide spread of separate/different domains such as marketing, information technologies, management science and finance. Studies conducted have found proof of a relation existing between technological progress and productivity in the banking sector. Moreover, it is emphasized that banks implement applied mathematics models that support their money knowledge for various functions. (Zhao, Xu and Kang) Comment by Author: Unclear Comment by Author: Of what? Comment by Author: Cite the sources that prove it!

These functions include credit evaluation and risk analysis. Financial sector reforms allowed a growth in competition and this led to bank lending morphing into a vital supply of funding. Credit risk analysis is by itself a vast domain that encompasses an oversized variety of analysis publications that exists within banking and have been unfolding though the last twelve years. Other banking- connected subjects where analysis has maintained a constant presence is fraud bar and detection within antiquated banking services as well as in new communication channels that support e-banking services. Comment by Author: This is the literature review. Where are the citations to the literature?

According to the International Journal of Science, Engineering and Technology Research, “E-banking is conjointly subject of another analysis domain associated with technology acceptance concerning new communication channels adopted by banks.” A recent team that has boomed upon further analysis is the drive by the global money crisis. Factors such as bankruptcy as well as connected subjects such as general risk and contagion play a huge role in this. Additionally, competition has play a huge role in the swaying of client connected areas with the banks in Nigeria that exploit Big data increasing their investment is the retention of clients, their customer relationship management and lastly their targeting. (Hassani, Saporta and Silva) Comment by Author: Of what?

Research in the field of banking is currently a noteworthy domain of analysis. Due to major advances in data technology, most if not all banking operations and procedures square measure automatic, generating huge amounts of information. Therefore, all the above mentioned subjects will most likely have the benefit of metallic element solutions. (Desai and Kulkarni) Comment by Author: unclear Comment by Author: unclear.

Banks in Nigeria that have implemented big data have been noted to have a certain edge above their competitors. The bank thrive of sentiment analytics. This is due to the fact that the banks have to continuously monitor what all their customers say for marketing purposes. Banks have to identify who the key customers are and through the acquiring of feedback they have the opportunity to improve all the loopholes in order to increase productivity and services thus providing a better product than before. (Desai and Kulkarni) Comment by Author: Source? Comment by Author: Not a sentence.

Moreover, the banks in Nigeria can also modify their service delivery. This is because big data would mainly comprise of a huge system although its job is to switch or shift tasks. Whenever any kind of reputation or account range is submitted into the system, it will sift through all the provided information and only provide the desired content. This can allow the banks to modify their work processes and will save them valuable time and prices. Large amounts of knowledge will enable organisations to point out and rectify problem areas before they have a lasting effect on their customers. Comment by Author: Source? Comment by Author: Source? You’re losing your connection to the literature you’re supposed to be reviewing.

One of the main problems faced by banks in Nigeria is fraud and its detection and prevention is a painstaking issue and many banks struggle with this but with the advent of big data in some, things are changing for the better. Big data ensures that no unauthorised transactions will be made therefore allowing for increased security and safety to the whole system. Comment by Author: Source? Comment by Author: Source? Comment by Author: Source?

Banks implementing big data have also enjoyed the benefit of enhanced reporting and concise delivery. This is due to the fact that the banks gain access to huge amounts of data that contain the needs of different customers and this allows them to offer those needs in a much more meaningful way. Through the use of this data, the banking industry in Nigeria can be able to provide exactly the information required by the customer instead of unnecessary data and filler content. (Alaraj and Abbod) Comment by Author: Big Data enables the customer to do this? Or, do you just mean a large database?

When it comes to risk management, big data plays a huge role in this. The early detection of fraudulent activities is a key part of risk management and huge amounts of information will do the maximum amount risk management due to the fact that it can be used for the identification of fraud. Massive amounts of information locate and present massive information on one large scale that easily produces it to cut back the amount of risks to a number that is manageable. This massive amount of information plays an integral role and the desegregation of the needs of a bank into a centralised and practical platform. Therefore, this reduces the banks possibilities of ignoring huge losses and terms of information and ignoring fraud. Comment by Author: Source? Comment by Author: Source? Comment by Author: Source? Comment by Author: Source? Comment by Author: Source? Comment by Author: Source?

In addition to this, customer segmentation can also be used in a greater and simpler fashion through big data. Big data allows for the conception of targeted marketing programs. This is done through the identification of card usage habits that the customer's portray as well as the creation of loyalty programs. Through the use of this way, a relationship between the bank and the valuable customers will be built. Furthermore, the bands can also be able to examine customer feedback to the collection of customer sentiment in text form from various websites on social media in. Once these sentiments have been collected in an orderly fashion, they can be grouped into good and bad or positive and negative categories. Then after the application of various filters they can be used to provide additional and improved services to the customers. (Desai and Kulkarni) Comment by Author: Source? Comment by Author: You don’t cite a great deal of literature. You keep using this one source over and over. Just discuss this source once and move to other sources.

Also, another benefit that can be exploited by the banks is the detection of a customer exiting the institution or facility. This is a huge factor due to the fact that the cost of acquiring new customers is much greater than the cost of retaining previous customers. If the bank takes care of understanding of the problem, necessary attention has to be done in order to find a solution. Comment by Author: Source?

Big knowledge analytics are currently being implemented across numerous of the banking sectors in western Africa, in Nigeria to be specific. This allows them to deliver high-quality services to the loyal customers both in an internal and external fashion. Additionally, it allows them to improve on their active and passive systems of security. (Zhao, Xu and Kang)

Data Analysis Comment by Author: You haven’t identified any sources of data. How can you analyse what you haven’t identified? I also don’t see the research questions that would structure the analysis.

There are several ways to implement the above-stated advantages into the banking system in Nigeria. The first factor that is to be studied is the monitoring of customer satisfaction with sentiment analysis. With the assistance of sentiment analysis, it is quite possible for one to understand the customer better and to deduce what the customer is saying concerning the bank in terms of service or product. Net bank Limited in South Africa as an example. This is because the bank realises the great advantage that they experience when they know what customers are saying about the banks through the aid of social media analytics. However, Nigerian banks usually make use of impractical methods that are not highly efficient because this forces there to read all the comments in a specific blog or review sites and then directly report the problem areas. This is highly impractical for them. (Alaraj and Abbod) Comment by Author: I thought the literature review was done and that you were turning to your own original research.

However, with the assistance of sentiment analysis or social media analytics the data which is in txt format can I love the use of a text analytical algorithm which is suitable for problems such as the one stated above. Naive Bayes algorithm is a good example of this. These search algorithms not only analyse text but they also produce highly negative to neutral and highly positive clusters or groups. Such a type of system can greatly assist the bank in determining if the customer is happy without reading all the comments are posted in social media such as blogs, review sites and popular platforms like twitter or Facebook.

Next, the bank could attempt to identify the key customers. If the bank has the ability to identify the key customers such as those who have been flagged for a high number of referrals, then the bank can prioritise them due to their importance to the institution because they are enjoying banking services and other products of the bank. These customers are more likely to spread confirmation to other customers and interact with them. Information concerning the bank will then spread and this will be good for the marketing of the institution with loyalty being bolstered at the same time. Such customers can easily be identified with assistance of link analysis which depends on highly unstructured social network data as well as data that is sourced from third party sources such as review sites, blogs, twitter and other popular sites such as Facebook. Furthermore, decision trees can also be used in the determination of how customers are relating with one another and then the tree can also be used to score customers from highest to lowest in a systematic ranking order. (Desai and Kulkarni) Comment by Author: Ibrahim, I’m going to stop reading here because I don’t see an original research paper anywhere. You don’t have research questions, methodology, data collection, data analysis, etc. There is nothing here that is publishable in a peer-reviewed journal. You have a great deal of work to do before the end of the semester.

In addition to this, the banks could also work on improving the customer feedback. if one examines the Nigerian banking industry with regards to how they treat customer feedback, one will deduce that they only user research tools such as customer surveys which are very time-consuming and also yield inaccurate results. The inaccuracy comes from the fact that the method only samples some groups not the entire population. However, if the bank decides to implement sentiment analysis tools then the system will be able to gather all of the customer reports on social media platforms as well as log and blogs. A good example of this is Barclays. This bank was able to compile data and form a report on their mobile application problems. The information was sourced from social media platforms in the form of analytics. After this, the bank was able to make several amendments in a prompt fashion. (Desai and Kulkarni)

Other successful examples of this include Dell which is a computer manufacturer. This company gathered reports from the World Wide Web concerning their products. Customers logged complaints about their devices overheating on social media platforms such as twitter and Facebook. Because of the social media analytic tools that the company uses they were able to know these problems almost immediately. If Nigerian banks can properly implement this method then they can be able to improve their packages and products through customer feedback.

In addition to this the Nigerian banks could also attempt to identify their customer profile. Typically Nigerian banks usually send a general mail or message to every customer but people view this message as nothing more than nuisance in current times and many just categorise it as spam. Most do not even read these messages. This is because some of the messages are not necessarily tailored for them, good example of this is messages detailing products for kids accounts sent to adults. (Hassani, Saporta and Silva)

Though, if the bank could use recommendations that can determine which project message is suitable to their customers through the use of a classification algorithm such as a decision tree or even a neural network this can assist them in the establishment of the proper content that their customers are interested in. The bank can also determine the products or services that the customer will engage in more through the use of cookies URL in order to determine which links were clicked most by the customer.

Furthermore, with big data the banks can be able to detect when the customer is about to go or exit. A huge problem that is being experienced by banks is the retaining of customers. A good example of this is union bank. this bank used to have the largest amount of customers in annual statistics that were released every year but over time as a result of general unawareness concerning the customers will exiting made lose a large percentage of their customer base. For banks to prevent this, it is necessary for them to understand their customer in a holistic way. Customer data transaction is a very effective way of deducing how a customer is banking. Sentiment analysis can also be used to flag if a customer is not satisfied with the institution and combined with survival analysis it is also the best way for want to know when a customer is leaving. The reason behind this is, survival analysis is adept at comparing different customer segments across a wide variety of time-series. (Zhao, Xu and Kang)

Also, customer segmentation is another factor that should be focused on. This is due to the fact that customer segmentation has evolved and is now beyond the traditional bounds of segmentation that group a customer base on their account type, marital status, age, account balance etc. Customer segmentation is moved beyond that and the current times necessitate a prediction system which can be able group the customer base on lifestyle, life stage and even special events/conditions such as when customers choose to buy online with debit cards or credit cards. This also includes the platform that the customers choose to shop with online. In order to successfully perform this advanced customer segmentation, a clustering algorithm should be used. (Alaraj and Abbod)

When it comes to security enhancement, the bank should prioritise this area. This is because data security is extremely crucial in the banking sector. Banks are currently implementing security protocols for their data and networks but it would be better for them if they had a certain level of awareness concerning the nature of the incoming threat. A good example of this is a certain company employing the use of datameer in order to track a virus that started in Russia and moved across the entire Asian continent and then made it to North America. If Nigerian banks to implement a system or product like datameer then they would benefit greatly improve the security in the process. Datameer could assist them in the flagging of viruses or threats that are in coming because the product follows the trend of how a certain virus is moving from a particular geographical area to another so that it can be able to predict the next move that the threat will make. (Noori)

Analysis and Algorithms

Banks can use some of these forms of analysis in order to either acquire data in a better fashion and also study it and break it down in a simpler more effective fashion. The types discussed above include:

· Survival analysis.

· Sentiment analysis.

In survival analysis, the things that one should consider when it comes to customer exits mainly is the customer transaction records, customer feedback and the customer relation records.

The following calculations can be used.

T is used to signify the response variable, T ≥0.

The survival function is S (t) = Pr (T > t) = 1−F (t).

The survival function is used to give the probability that a subject will survive past time t.

As t ranges from 0 to ∞, the survival function has the following properties:

· It is non-increasing at time t = 0, S (t) = 1.

In other words, the probability of surviving past time 0 is 1. At time t = ∞, S (t) = S (∞) = 0. As time goes to infinity, the survival curve goes to 0. – In theory, the survival function is smooth. In practice, we observe events on a discrete time scale (days, weeks, etc.).

The hazard function, h (t), is the instantaneous rate at which events occur, given no previous events. h(t) = lim Δt→0 Pr(t < T ≤ t +Δt|T > t)/ Δt= f(t) S(t). The cumulative hazard describes the accumulated risk up to time t, H(t) =Rt 0 h(u)du.

On the other hand, sentiment analysis is an algorithm that can be used to analyze the sentiment of the content on social media platforms such as Facebook and Twitter. Content like tweets as well as status updates can be used to return sentiment rating for negative, neutral and positive. All the text will be extracted through text analytics and then it will go through all of the content (tweets and status updates) that is produced by the customers. The analysis run will return the resulting rating in the form of positive, neutral and negative.

The algorithms that can be used include:

· Classification algorithm.

· Clustering algorithm.

A classification algorithm or a neural network is a classification type algorithm. This means that it assigns data to a target field that has been predetermined. This target field would essentially be a form of scoring function that could be a customer’s propensity towards a certain type of product or even the estimated value that a house possesses. (Widrow, Rumelhart and Lehr)

A clustering algorithm employs the use of cluster detection which assists by the detection of clusters of data that form natural groups that exist within the data set. A good example of this is when a cluster could contain grouping of a customer base on life stage and life style. The steps of the above stated algorithm are outlined below:

· The selection of random K-points that are seeds for the centroids of the K-clusters.

· The assignment of each point to the centroids that are closest to the point.

· After all the points have been assigned then the recalculate new centroids of each and every cluster.

· The 2nd and 3rd steps will then be repeated until all the centroids no longer move.

Additionally, datameer can also be used by the Nigerian banks. This is an end-to-end platform for big data that ignores the limitation of ETL as well as static schemas in order to empower the bank in order to integrate data acquired from any source into Hadoop. This product has pre-built data connector wizards that are used for common structured, unstructured and semi-structured data sources. This method greatly simplifies data integration. It also ensures that the users are usually up to date and it provides the link any of the other sources. The data integration process is also quite rapid and simple.

Recommendation

With the rapid growth in mobile use and the accessibility of social media through smartphones, its critical banks continue to transform by using data and insights to connect with their clients. Across the world, financial institutions are investing heavily in mobile-friendly apps that make it easy for clients to conduct banking transactions while on the go. The result is rewarding – adds a deeper, value-added service for bank clients and provides a competitive edge for banks.

Banks are also continuing to look ahead at future innovations and technology advances to strengthen security, regulation compliance and improve fraud detection and prevention.

Big Data gives companies an incredible volume of information and records at your disposal. Where banks and financial institutions might have tracked performance and metrics on a branch-by-branch or region-by-region basis in the past, the more advanced companies now have access to all their records, simultaneously. (Quinlan)

That means they can build more accurate models of customer behaviour and set proper pricing for loans and financial products to optimize profits and align the right products with the right customers. Multi-variant data analytics is key in decision making and particularly when you are testing a new hypothesis. Big Data can also help define a baseline for ‘normal’ operations, which gives companies a head start in detecting fraud and helps manager’s spot compliance and regulatory issues before they become a problem.

Banks have become basically software shops, employing thousands of programmers. But, while start-ups with only a handful of programmers can crank out new, innovative services, banks are unable to, despite having more resources. Banks need to become more agile, quick and innovative as it pertains to software, just in order to keep up.

A top trend in Big Data technologies in the banking industry will be to better read and understand Big Data. They need to optimize the maintenance of legacy applications and use the freed up capacity for new, innovative stuff. Given the massive amount of programmers they have, this would be a huge force and they could change the game to their advantage. (Desai and Kulkarni)

Conclusion

The technical key to the successful implementation and exploitation of Big Data and digitization of business processes is the ability of the organization to collect and process all the required data, and to inject this data into its business processes in real-time - or more accurately, in right-time. Big knowledge analytics is currently being enforced across numerous spheres of banking sector, and helps them deliver higher services to their customers, each internal and external, alongside that is additionally serving to them improve on their active and passive security systems. If the banking industry can implement all the above analysis and algorithm effectively (sentiment, clustering, link, survival, decision tree or neural analysis and datameer) it will increase the efficiency of the banking system.

References

Alaraj, M. and M.F. Abbod. "Classifiers consensus system approach for credit scoring." Knowledge Based Systems (2016): 89-105. Chunsheng Zhu, Huan Zhou, Victor C.M. Leung, Kun Wang, Yan Zhang, and Laurence T. Yang. "Toward Big Data in Green City." IEEE Communications Magazine (2017). Desai, D.B. and R.V. Kulkarni. "A Review: Application of data mining tools in CRM for selected banks." Int. J.Comput. Sci. Inf. Technol. (2013): 199-201. Hassani, H., G. Saporta and E.S. Silva. "Data Mining and Official Statistics: The past, the present and the future." Big Data (2014): 34-43. Noori, B. An Analysis of Mobile Banking User Behavior Using Customer Segmentation. 2015. Quinlan, J.R. C4.5: Program for Machine Learning. Burlington, MA: Morgan Kaufmann, 1992. Widrow, B., D.E. Rumelhart and M.A. Lehr. Neural networks: Applications in industry, business and science. ACM, 1994. Zhao, Z., et al. "Investigation and improvement of multi-layer." Expert System Applications (2015): 3508-3516.