Technology developments in artificial intelligence (AI) are extremely active. Be-hind the ethical controversy of generative AI in society, a shift in AI frameworks is taking place and it attracts attention in open source software (OSS) community. This research proposes a methodology to identify the next OSS in order to make data-driven decisions about OSS, using data from AI frameworks. The method-ology identifies OSS repositories whose development rises and falls are linked, not only by correlation, but also by taking the contributors involved in the two repositories as causal relationships. In this research, a relationship that rises and falls together is called complementary OSS and a relationship that rises and falls in opposite directions is called substitute OSS. As a result, It turns out that both PyTorch and TensorFlow continue to grow together. Pieces of supportive OSS were identified for both complementary and substitute OSS, rather than OSS that were complete replacements. This will contribute to identifying the next OSS with higher accuracy.