SenteraLLC/geoml

Issues

• Define "feature groups" for training

  #9 opened 5 years ago by tnigon
  1

• Weather DB Schema Plans (long term storage v. API call)

  #24 opened 5 years ago by tnigon
  0

• `SentinelSat` Install Bug

  #49 opened 3 years ago by bohm0072
  1

• Create `Visualize` class for generating GeoML output graphics

  #51 opened 4 years ago by tnigon
  0

• Generalize FeatureData so it works for any kind of data

  #34 opened 4 years ago by tnigon
  1

• Clean up "tune" vs "train" variable/function names in Tuning class

  #35 opened 5 years ago by tnigon
  0

• Integrate with `db` API so training data can be retrieved from Postgresql

  #38 opened 4 years ago by tnigon
  1

• Update FeatureData to support changes to data sources

  #41 opened 4 years ago by tnigon
  1

• Predict class for making geospatial predictions on new data

  #46 opened 4 years ago by tnigon
  2

• Set up CircleCI for continuous integration for new code contributions to github

  #29 opened 5 years ago by tnigon
  2

• Wrap config.py config.config_dict in a class so values don't change

  #32 opened 5 years ago by tnigon
  0

• Add installation/setting up environment in README

  #27 opened 5 years ago by tnigon
  1

• Create Training function to train based on all data

  #25 opened 5 years ago by tnigon
  1

• Evaluate Sentinelsat and Sentinelhub libraries to access satellite data

  #26 opened 4 years ago by tnigon
  1

• Create class to deal with feature selection

  #16 opened 5 years ago by tnigon
  5

• Design inherited classes so they can accept a "parameter dictionary" that contains all the information needed to execute

  #17 opened 5 years ago by tnigon
  2

• Create a utilities class that holds code that is re-used across classes

  #18 opened 5 years ago by tnigon
  2

• Tuning.regressor_name include transform/pipeline information

  #20 opened 5 years ago by tnigon
  3

• Create class diagram of research_tools package

  #21 opened 5 years ago by tnigon
  1

• Split train and test sets before putting into matrix

  #11 opened 5 years ago by tnigon
  1

• Add function to `feature_data` to get or set the random seed

  #12 opened 5 years ago by tnigon
  1

• Have option to save X and y data to file in `get_feat_group_X_y()` function

  #13 opened 5 years ago by tnigon
  1

• Create `Training` class that inherits from FeatureSelection

  #14 opened 5 years ago by tnigon
  4

• Add `get_repeated_stratified_kfold()` and `check_stratified_proportions()` to `feature_data` class

  #15 opened 5 years ago by tnigon
  1

• Set up basic testing for every class in research_tools

  #6 opened 5 years ago by tnigon
  1

• Use skopt.gp_minimize for hyperparameter tuning optimization

  #7 opened 5 years ago by tnigon
  0

• Design an inheritance architecture for all data, and tuning and training functions

  #8 opened 5 years ago by tnigon
  5

• Create a dictionary describing the data to include in the X matrix of every "feature group"

  #10 opened 5 years ago by tnigon
  2

• Create class object for joining data

  #1 opened 5 years ago by tnigon
  1

• Create setup file and __init__.py

  #2 opened 5 years ago by tnigon
  2

• Create file management class that returns full X matrix for any feature group

  #3 opened 5 years ago by tnigon
  2

• add join_closest_date() function to join_tables object

  #4 opened 5 years ago by tnigon
  1

• join_tables add a function to add/change ``base_dir``

  #5 opened 5 years ago by tnigon
  1

• Change name from research_tools to geoml

  #48 opened 4 years ago by tnigon
  0

• Decide how to handle multiple types of spectral data

  #42 opened 4 years ago by tnigon
  0

• Weather data - use 3rd party API to populate DB table for every point

  #23 opened 5 years ago by tnigon
  0

• Make a plan for using NewtDB (an extension of ZODB for PostgreSQL)

  #22 opened 5 years ago by tnigon
  0

• Lasso: check difference between normalize=True/False

  #19 opened 5 years ago by tnigon
  0