Just jotting down what we may need in the future

[junipertcy]

• More user friendly Dandi io (summary of metadata, etc.)
• At least two State-of-the-art solvers (QUIC, pISTA or gISTA)
• Module that generate synthetic graphs
  • random
  • core-periphery
  • planted-partition
  • k-core
• Module that computes a graph's summary statistics
  • mean degree
  • Betti-1 and Betti-2 (treated as a flag complex)
  • "core-ness" (what are these?)
  • Perhaps consider DomiRank?
  • Others??
• Module for testing the robustness of results
  • e.g., which binning size is robust? This may be different across behaviors. #2
  • what regularization parameter to use?
  • Bootstrapping for one mouse's data (to test the robustness of single experiment result)
  • Need to be careful when binning in between behaviors (re-scale, discard, etc)
  • Need a function to throw away low-excitation neurons