Abstract: Since data was officially regarded as a production factor, the issue of opening public data has been becoming more and more important. This paper acquires policy texts of opening public data in multiple places by means of Python programming, then analyzes their publishing time, quantity and geographical distribution characteristics, and uses word frequency statistics and text similarity calculation, to study the overall situation and existing problems so as to provide suggestions and reference for the formulation of relevant policy texts. Firstly, it investigates the definition of public data and builds the expression framework; secondly, it discusses whether public data should be opening to everyone; thirdly, it expands the tool dimensions of government security policy for the subsequent security issue, and explores the
evolution trend.

Key words: opening public data, Python, word frequency statistics, text similarity