Recently, Taiwan government has voluntarily opened up various data sets to provide the public for analysis and application in order to find potential problems and help solve some public issues. A common way of using the data sets to initially explore potential issues is to visualize the miscellaneous data and present useful information represented by readily available statistical charts. In this process, the most often encountered problem is the fluctuations of the data changes in the time series of the statistical chart, so it is not easy to visually compare the trends of the source data among different factors. Therefore, we use the concept of moving average to smooth short-term fluctuations. The moving average can clearly show the difference or impact of the data on different factors. However, when using some of the traditional moving average methods to draw charts, we found that the moving averages lag behind the values of the source data, making the data visual delay in comparison. To solve this problem, this study proposes two improved moving average methods, which are compared with the traditional moving average methods in terms of smoothness and lag results to see if we can find a better one, which can more accurately analyze time series datasets to determine the trends of the past problems and the possible future trends, and then lead the follow-up study.
Relation:
International Journal of Electronics and Information Engineering
