• Setup Python package for the bttv i.e. Bradley-Terry Time Varying model. This is done here
• Add submods (git submodules) and clone the Python package
• Setup reproducible workflow for our package development

• make it working in three different options

• give an option for controlling the verbose level

• give an option for controlling output direction

• rerun NFL data using cross-validation

mainly about "mu_parameters" in beta_gaussian_process

• elementary functions, including the objective, gradient, proximal operators, et al.

• functions implementing an optimization method

• numpy should be imported in this module
• loglike could be confusing as it actually returns negative loglikelihood (suggestion: neglogike)

• Delete python/old_backup folder
• Move jupyter notebooks to top level reproducible_analysis directory
• Delete unused jupyter notebooks
• Number remaining jupyter notebooks in a workflow
• Update wiki with numbered jupyter analysis

Basic calculations

Todo list

• get proximal operator for L2-penalty
• get proximal operator for L1-penalty
• implement proximal gradient method
• implement backtracking line search
• implement acceleration

• crude analysis of NASCAR, squared-l2

• crude analysis of NFL, squared-l2/l2

• check indices of teams in NFL data

• more methods of optimization

• detailed analysis/interpretation of results

• Add list of papers
• Add data directory
• Add python directory
• Add R directory
• Add presentations directory
• Add report directory
• Add some papers e.g. conjugate gradient descent to the papers directory
• Add README md file
• Add CONTRIBUTING md file

Todo list

• decide if the remained data guarantee the existence of solutions
• decide if the remained data guarantee the uniqueness of solutions
• implement cross validation

• Delete presentations directory (will create a posters directory later)
• Update conda environments for pytest
• Delete papers directory (will create a posters directory later)
• Create separate CONTRIBUTING wiki page - link to README
• Create separate SETUP wiki page - link to README
• Update Makefile for the following
  • Get NFL data
  • Get NFL ELO data
  • Get NASCAR data
• Number jupyter notebooks
• Have references to charts in jupyter notebooks

• Update NASCAR race rankings to download for all races
• Setup jupyter notebook to read in data

• Add nascar dataset and setup
• Add NFL data from nflscrapR and setup.
  • Per discussion with Ron this is a good reference and this one

• Clean up report dir and add only pdf paper
• Create nascar data directory
• Remove old nfl data

Including beta and game records

• Update bttv python package
• Update references to bttv