Collection of Python functions to run network-based analysis in signed and directed networks. Follow the jupyter notebook tutorial to see what you can do. The notebook jumps between python and R. Beside the packages below, you will also need to install
Developed on python 3.7.3 installed through ANACONDA [Clang 4.0.1 (tags/RELEASE_401/final)] in macOS Catalina version 10.15.1.
| numpy | 1.18.1 |
| pandas | 1.0.1 |
| networkx | 2.4 |
| matplotlib | 3.1.3 |
1. Weight and sign are collapse in the same value: weight. Negative weights indicate inhibition, while positive one are activations.
2. We use directed graphs to calculate network and node-based measurements. This type of graphs only allow ONE interaction between two nodes. As the weight contains also de sign, if an interaction is repeated with opposite signs between 2 nodes (e.g. A -| B; A -> B), only one weight can be kept. Thus, the weight that is kept is the highest (keeping the sign).
The function __get_measurments__ calculate the follow network's features: - Number of edges - Number of nodes - Density - Avg betweenness centrality - Avg degree centrality - Avg eigenvector centrality