Excel-based complaint data is multivariate data such as report date, product name, customer name, defect details, defect cause, responsible department, supplier, countermeasures, shipping date, and cost, but there is a trick to making problems visible with a graph based on this. In a graph with the year on the X-axis and the number of defects on the Y-axis, divide the Y-axis by product, responsible department, defect details, etc. to see the problems. In a graph with the defect details and suppliers on the X-axis in order of frequency, divide the Y-axis by year to see the improvement trend. By arranging the number of days to failure (date of occurrence - date of shipment) on the Y-axis, the trend in the number of days to failure can be seen. By arranging the number of days to failure (date of occurrence - date of shipment) on the Y-axis, the trend in the number of days to failure can be seen. In this way, by dividing the X-axis or Y-axis or overlapping with colors, you can arrange the third and subsequent variables to see the problems. It is a good idea to narrow down variables with long tails to the main level with a data filter. If you find a singularity, select it and drill down with a subset. A bar graph looks good, but it is not suitable for discovering such problems.
