When I write my thesis, I mostly use economic data of provincial or prefecture-level cities. These data are generally obtained from yearbooks, statistical bulletins of national economic and social development, various statistical annual reports or government websites, but I always encounter the problem that incomplete statistics lead to missing data. For example, environmental pollution related indicators, the number of patent applications granted, etc. For these missing data, my method is generally as follows:

This method is generally applicable to: unilateral missing (usually missing in previous years, and the data in the latest year is generally complete).

A teacher told me that the missing values calculated by cubic spline interpolation method are smoother and the results may be relatively reasonable. It can be realized by looking up matlab, but it can only be actually operated when the data is processed next time.