e-CSTI

1. Purpose of “visualization”

Each government ministry and agency creates standardized review sheets (hereafter, “administrative project review sheets”) for all projects under their management (approximately 5,000 projects), to disseminate comprehensive updates on the implementation and funding of these projects. Administrative project review sheets from all ministries and agencies were collated in a database over a four-year period since the beginning of the 5th Science and Technology Basic Plan in 2016. By adding the calculated similarities between each entry of the administrative project review sheets and major policy issues in national plans (the 5th Science and Technology Basic Plan and Integrated Innovation Strategy 2019), sorting functionality of the content of the review sheets according to similarities are realized. This search functionality is not limited to science and technology budgets, but includes all administrative projects for which an administrative project review sheet has been prepared.

This “visualization” shows lists of administrative projects with a strong degree of similarity to specific policy issues and fields. For example, it enables efficient searches of changes in annual budget amounts per policy issue, and confirmation of whether similar administrative projects are being implemented by different ministries and agencies.

There are two visualization functions for searching administrative project review sheets: 1) dealing with initiatives tied to the 5th Science and Technology Basic Plan, and 2) dealing with key fields targeted in the Integrated Innovation Strategy 2019.

2. “Visualization” method

Administrative project review sheets were obtained from public data prepared by the Headquarters for the Promotion of Administrative Reform Cabinet Secretariat. Information shown on review sheets from the 5th Science and Technology Basic Plan period (between 2016 and 2019) was transferred into a database. Since the policy issues indicated in the 5th Science and Technology Basic Plan have a hierarchical structure, they were divided into intermediate sections or subsections according to the sectional structure of Chapters 2 through 7 and the Table of Contents. Finally, intermediate sections and subsections were defined as policy issues (64 items in total). Similarly, a total of 11 items from the Integrated Innovation Strategy 2019, including the six items from Sections (1) to (6) in Chapter 5, “Major Fields That Required Increased Effort,” and the five items indicated in intermediate sections 1 to 5 of Section (7), “Other Major Fields for Realizing Integrated Innovation,” were used as search target fields.The degree of similarity between review sheets and policy issues or fields is determined by using tf-idf (tf: terms frequency, idf: inverse document frequency to calculate feature vectors for individual nouns extracted from each sentence explaining the 64 policy issue items in the 5th Science and Technology Basic Plan, or each sentence explaining the 11 fields in the Integrated Innovation Strategy 2019, as well as explanatory sentences in project summaries from each administrative project review sheet; and the cosine similarity between these feature vectors is then used as the degree of similarity. After this, numerical  values, Japanese era names, generally suggested nouns, and other common nouns are excluded from calculations as stop words. The results from the degree of similarity calculations are displayed in the second “degree of similarity” column on the left in the list of administrative projects at the bottom of the visual display. It should be noted that the degree of similarity in these calculations does not account for semantics.

3. “Visualization” results

Below is an example of a visualization of the degree of similarity with policy matters in the 5th Basic Plan. Similarity between policy issue sentences is calculated, and then administrative projects with summaries that have a high similarity with the selected policy issues are listed. A bar graph shows the percentage of executed budgets per managing ministry and agency. These may be narrowed down with visualization tools according to the list and additional conditions.

Below is an example of a visualization of the degree of similarity with the Integrated Innovation Strategy 2019. Similar to the 5th Basic Plan, similarity is calculated between policy issue sentences, and then administrative projects with high similarity are listed. A bar graph shows the percentage of executed budget amounts per managing ministry and agency. These may be narrowed down with visualization tools according to the list and additional conditions.