- Folder IO:
- from_file
- from_postgis
- Folder graph preprocessing —> from tracking data to graph:
- activity graph
- graph_utils (only utilities for the preprocessing)
- Folder analysis:
- graph_features (maybe divide into single-value features and _helper like degree - distribution)
- derive_subgraphs (copy from current utils.py)
- graph_matching (feature-based similarity comparison of two graphs)
- graph_hypothesis_testing (from my network analysis paper)
- Folder visualisation
- ActivityGraph is the main class --> implement as in Trackintel that everythign is in separate folders and AG class gets important methods as wrappers
- The other functions are implemented such that the first argument is a graph, e.g. node_degree(AG, ...)
- Write wrapper functions in AG class for the methods that are in other files (e.g. IO files)
Important: Examples