A novel estimation of scientific and technological innovation index using multi-modal data

Scientific and technological innovation (STI) index refers to the level of STI and has become a powerful indicator of national competitiveness, since STI can help revolutionize traditional products and generate new products, and lead to rapid economic growth, and obvious energy conservation. Traditionally, assessing the STI index has mainly relied on statistical data, which has a coarse spatio-temporal resolution. Thus, this study proposes a novel method to estimate the annual and monthly STI index at county level by fusing multi-modal data: NTL data and POI data. This method demonstrates robust performance with an <i>R</i>² value of 0.81 and has a strong generalization capability. By analyzing the driving factors under different innovation development patterns, we conclude strategies to improve the STI level in counties following various patterns of STI. Due to its accuracy and effectiveness, our method has potential for application across a wide range of regions and over extended time periods.