Fuzzy Time Series and Data Visualization for Forecasting Sales of Grocery Ingredients

The development of information technology is currently increasingly rapid and produces very large amounts of data. Grocery store or also commonly known as UD. Ridho has several problems in determining the availability of basic food ingredients and the amount of demand for goods is uncertain, which means that grocery stores often experience shortages of supplies and/or excess supplies of basic food supplies at certain times. The forecasting method that will be used in this research is Fuzzy Time Series (FTS) with the aim of overcoming this problem by estimating sales of goods (basic necessities) at UD. Ridho in Dolok Masihul. The results obtained are a system that was built to predict sales of basic food ingredients at UD. Ridho in Dolok Masihul is suitable for use so it is hoped that it can help UD owners. Ridho in determining the amount of basic food supplies in the future correctly so that UD. Ridho did not experience any losses or reduced income.


Introduction
The development of information technology is currently increasingly rapid and produces very large amounts of data.Therefore, an effective method is needed to analyze and utilize this data so that the right decisions can be made.One method that can be used is Fuzzy Time Series (FTS), which is a data processing method that can overcome uncertainty in data and predict future values.Apart from that, data visualization can also be used to help understand patterns and trends in sales of basic food products.
A grocery grocery store is a type of business that sells various kinds of basic food items in large quantities (Fadilah Sari, 2021).
In determining the availability of basic food ingredients, it is often hampered by the uncertain amount of demand for goods, which means that grocery stores often experience shortages of supplies and/or excess supplies of basic food supplies at certain times.This problem was also felt by one of the UD grocery stores.Ridho which is located in Dolok Masihul.Grocery store or also commonly known as UD.Ridho has several problems similar to the problems above.Problems often experienced by UD owners.Piksel 11 (2): 425-434 (September 2023) Ridho is an excess or shortage of basic food supplies at a certain time which causes significant losses and reduced income.Given these problems, careful planning is needed to estimate the supply of basic necessities in a way that does not result in losses and reduces UD's income.Ridho.
For this reason, a forecast is needed that can provide an overview of the process of forecasting the number of sales of basic food items, so that it is hoped that it can provide information regarding forecasting the number of basic food materials to shop owners to prepare for better availability of goods and create innovations and good sales marketing strategies (Bayu Elfajar et al., 2017) (Wulandari, 2020).
The forecasting method that will be used in this research is Fuzzy Time Series (FTS) with the aim of overcoming this problem by estimating sales of goods (basic necessities) at UD. Ridho in Dolok Masihul.This method was chosen because the FTS method has a good level of accuracy and can be combined with other approaches in solving forecasting problems (Nugroho, 2016).With this research it is hoped that it can help UD owners Ridho in determining the amount of food supplies in the future so that UD.Ridho did not experience any losses or reduced income.

Research Method
In this methodology, the author conducts research using two techniques, namely quantitative and qualitative.Quantitative techniques involve analyzing historical data and producing numbers in the form of time series data.Data collection can be done through primary sources (directly from data providers) and secondary sources (through other people or documents).Meanwhile, qualitative techniques involve subjective opinions that cannot be expressed in the form of numbers or values.
In the sales forecasting system developed in this study using the Fuzzy Time Series method and data visualization applied to the calculation of forecasting values.The flow of calculations using the fuzzy time series method where the data needed is historical data from previous years.The data is used as a reference for forecasting raw material requirements data in the following month by determining the number and length of intervals, after calculating the interval length data, checking the calculated data is carried out, if the data does not match then the process will be repeated from the start of inputting sales data.However, if the data matches then the process will continue to the Fuzzyfication section.Then from these results, a fuzzy logical relationship is determined and eliminated with fuzzy logical relationship groups to find forecasting data.At this stage, the authors make predictions on the data that has been collected in the second stage using the fuzzy time series method.As for the steps of the fuzzy time Next is to determine the set of Fuzzy Logic Relationships (FLR).The following is the FLR for the data on sales of staple foods in the "Rice" category: The final step is to carry out the Defuzzification and Forecast calculation process.The following are the forecasting results for each category of basic food sales at UD. Ridho in Dolok Masihul is shown in table 4.

System Interface Page Design
The process in an implementation on a website application.the interface design will display the structure or commonly referred to as a sketch based on the application of the application of fuzzy time series with data visualization for the sale of UD basic food ingredients.Ridho in Dolok Masihul.The fuzzy time series results page is an implementation display for the fuzzy time series forecasting results and data visualization that you want to do or who want to see the forecast results.The fuzzy time series results page can be seen in Figure 2.

No
The

System Implementation
At this stage, the results of the system implementation that have been carried out based on the analysis and design of the system that have been previously designed will be described.This stage contains the stages of system development using the Fuzzy Time Series method and data visualization for forecasting sales of UD basic food supplies.Ridho at Dolok Masihul.page will display information in the form of forecasting results, calculation of mean absolute percentage error & visualization of data that was successfully formed in each period and type of basic food ingredients.

Discussion
The

Find
Figure 1.Fuzzy Time Series Flowchart

Figure 3
Figure3shows the Forecasting Results page, calculation of mean absolute percentage error & visualization of the data that will be displayed to the user.This Fuzzy Time Series and Data Visualization for Forecasting Sales ofGrocery Ingredients PIKSEL status is accredited by theDirectorate General of Research Strengthening and Development No. 225/E/KPT/2022 with Indonesian Scientific Index (SINTA) journal-level of S3, starting from Volume 10 (1) 2022 to Volume 14 (2) 2026.427

Table 2 .
Fuzzy Logic Relationship (FLR) Data on Sales of Rice Basic Foods

Table 3 .
Defuzzification of Sales Data for Basic Food Materials, Rice

Table 4 .
forecasting results for each category of basic food ingredients

name of the Groceries Category Forecasting Results
Source: Research Result (2023) Figure 2. Fuzzy time series forecasting results calculation page test that has been carried out in this study is by testing the percentage error rate using MAPE.Testing the percentage error rate using MAPE aims to determine the error rate from the forecasting results produced by the system in forecasting sales of basic food ingredients at UD. Ridho in Dolok Masihul.Testing the percentage error rate in the application of the Fuzzy Time Series and Data Visualization methods in this study was carried out for each category of basic food ingredients, namely rice, flour, oil, detergent, grains, instant noodles, spices & cooking ingredients, packaged drinks, snacks, soap & shampoo, cigarettes.The test is based on the resulting MAPE value as a reference for the error rate of forecasting sales of basic food ingredients at UD. Ridho in Dolok Masihul.
The following is a recapitulation of forecasting results and Mean Absolute Percentage Error (MAPE) sales of basic food materials at UD. Ridho in Dolok Masihul is shown in table 5.

Table 5 .
Recapitulation of forecasting results the amount of basic food supplies in the future correctly so that UD.Ridho did not experience losses and reduced income.NoThe

name of the Groceries Category Forecasting Results
of proposal seminar exams, and professional colloquium exams, All Lecturers and the staff of the Computer Science Department who provided very useful knowledge and were indirectly involved in the preparation of this Final Assignment, the Author, his Parents and all his beloved family who were the main motivation for the author in completing the preparation of this Final Assignment and have given many prayers and support for authors, Shop owners UD.Ridho who has provided and permitted to carry out research, all friends who have provided encouragement, support and are useful to the author when experiencing difficulties. series