Shulou(Shulou.com)06/01 Report--

In this issue, the editor will bring you about how to use big data technology to optimize ATM operation. The article is rich in content and analyzes and describes for you from a professional point of view. I hope you can get something after reading this article.

Banks need huge investment to maintain the operation of various businesses every year. How to effectively improve bank operation efficiency and reduce operating costs has always been an important goal for banks to pursue. At present, many business operations of banks still rely on specific responsible persons to "clap their heads" on the basis of experience and intuition to make decisions. Big data's analysis can find the operation rules from the data and provide a decision-making basis for operation optimization. Below, we will take several applications of big data Technology to optimize ATM operations as examples to illustrate how big data optimizes operational efficiency and reduces operating costs.

1. Big data Technology optimizes ATM Operation

1) Optimization of ATM configuration (location and type)

Through some simple statistical analysis of the log data of ATM machines, we can help to improve the configuration of ATM to better serve customers, improve service quality and reduce costs. As shown in figure 1, by doing a statistics on the historical transaction volume of ATM and showing it in a bar chart, we can find that some ATM usage is very low (for example, the ATM15 usage in the figure is almost half that of ATM1). It shows that the location of some ATM is not reasonable. Combined with the geographical location information of ATM, we can relocate and relocate the low-usage ATM to improve the use efficiency and reduce the bank cost.

Figure 1. ATM transaction count

For the transaction analysis in figure 1, if the granularity of our analysis is finer, for example, if we analyze the transaction records of each ATM machine by transaction type, we can get the distribution map of the transaction type as shown in figure 2. Through the pie chart display, we can find that the distribution of various types of ATM2 transactions shown in the left chart is relatively uniform, but the ATM5 shown in the right chart mainly includes only four transaction types, of which deposit transactions are very few, which shows that people near this ATM have less demand for deposit business. This all-in-one machine can be replaced by a cash machine, thus reducing bank costs.

Figure 2. Analysis of ATM transaction types

2) optimize the ATM menu to reduce the waiting time

In addition, reducing the waiting time of ATM queue is an effective measure to improve customer satisfaction and loyalty. By analyzing ATM log records, we can understand the menu paths of customers using ATM machines and which operations they spend a long time on, so we can use the results of data analysis to optimize ATM menu design and shorten menu paths, so as to improve the efficiency of user operations. For example, most ATM now have the function of quick withdrawal (FastCash), but the design of the amount of quick withdrawal may not be consistent with the actual demand. Through the statistical analysis of the data, we can understand the most commonly used amount of withdrawal and optimize the menu. In addition, personalized menus can be provided based on the results of data analysis to meet the needs of different user groups.

Figure 3. Example of ATM Fast Cash function

3) forecasting the cash demand of ATM

In addition to statistical analysis of ATM data such as 1Magazine 2 to support operational optimization, we can also optimize operations by deep mining of data and building machine learning models. For example, the cash management of ATM is a very important operation management part of the bank, which directly affects the normal operation and profit of the banking business. It is an operating cost for banks to store too much cash in ATM, but it is easy to be empty if there is too little cash, which reduces the quality of service and needs to increase the operating cost of armored vehicles.

By forecasting the demand for cash in advance, we can increase banknotes on demand and reduce the cost of banks. However, the demand for cash has great uncertainty, which is related to many factors. Taking the time factor as an example (as shown in figure 4), the ATM withdrawal operation reaches its peak before holidays and weekends, and time series analysis is needed to forecast ATM cash demand, and the weekly law, monthly law and holiday law are found out from the data. In order to accurately predict the cash demand, it is necessary to comprehensively consider a variety of possible related factors and build a complex machine learning model.

Figure 4. ATM withdrawal amount time series

4) ATM predictive maintenance

Another important application of big data analysis to ATM data is predictive maintenance. ATM downtime will not only bring business losses to banks, but also give users a bad impression and reduce user loyalty.

Figure 5 ATM downtime example

At present, most banks still use the traditional break-fix method for the maintenance of ATM, that is, as shown under the dotted line in figure 6, the failure occurs first and then maintenance is carried out. By building a machine learning model based on historical ATM data, we achieve predictive maintenance (predict-and-prevent), predict the downtime of ATM in advance, and then schedule maintenance during off-peak hours, so as to avoid business losses caused by maintenance and improve user satisfaction.

Figure 6. ATM Predictive maintenance Framework

two。 Landing case of operation optimization

1) DBS Bank of Singapore [1]

DBS Bank can translate useful ATM usage data and customer behavior data into day-to-day implementation plans to help banks work out plans to arrange banknotes during peak hours. Through the use of the new solution, the shortage of banknotes in machines has been reduced by 80% (that is, empty ATM machines), more than 30,000 hours of customer waiting time has been saved, the amount of cash left in the bank has been reduced by 40%, and the required banknote transportation arrangements have been reduced by 20%.

2) A bank in Australia uses big data to optimize the layout of its outlets [2]

Combining the internal data of the bank (including the distribution and performance of existing outlets, etc.) and external data (such as population, demographic structure, income level, etc.), the bank assessed 350 regions and optimized the layout of bank outlets.

3) Minsheng Bank [3]

Minsheng Bank uses big data's analysis to support the selection of outlets and ATM cash management.

4) A foreign commercial bank

The bank carries out predictive maintenance of ATM machines based on ATM data analysis, effectively reducing downtime.

Through data analysis to assist and support operational optimization decisions, we can avoid top-to-bottom "slapping on the head" decision-making, effectively improve operational efficiency and reduce operating costs.

The above is the editor for you to share how to use big data technology to optimize ATM operation, if you happen to have similar doubts, you might as well refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.