TODO write readme...

Until then here is an example:

this is an example of data to be fed into the DataLook object

stack = {'datasets': [ {'name': 'insurance_rates', 'type': 'csv', 'path': 'insurance.csv', 'y_values': ['charges'], 'drop': []}, {'name': 'heart_disease', 'type': 'csv', 'path': 'heart.csv', 'y_values': ['target'], 'drop': []}, {'name': 'IBM_hr', 'type': 'csv', 'path': 'IBM_hr.csv', 'y_values': ['Attrition'], 'drop': []}, ], 'transforms': [{'fname': 'self.categorize', 'params': {}}, ], 'models': [{'fname' : 'RandomForestRegressor', 'params' : {}}, {'fname' : 'DecisionTreeRegressor', 'params' : {}}, {'fname' : 'Lasso', 'params' : {}}, {'fname' : 'ElasticNet', 'params' : {}}, {'fname' : 'SVR', 'params' : {'kernel' : 'rbf'}}, {'fname' : 'SVR', 'params' : {'kernel' : 'linear'}}, {'fname' : 'SGDRegressor', 'params' : {}}, {'fname' : 'Ridge', 'params' : {}}, {'fname' : 'ExtraTreesRegressor', 'params' : {}}, {'fname' : 'AdaBoostRegressor', 'params' : {}}, {'fname' : 'GradientBoostingRegressor', 'params' : {}}, {'fname' : 'HistGradientBoostingRegressor', 'params' : {}}, {'fname' : 'RandomForestRegressor', 'params' : {'max_leaf_nodes' : 50}}, ], 'metrics': ['r2_score', 'explained_variance_score', 'median_absolute_error', ]}

create a DataLook object

dl = DataLook()

clear anything that exists - this is helpful in a notebook to not have to reload

the stack clears anytime it is run but jis removes any unrun data

dl.clear_stack()

the resutls continue to stay in a dataframe until cleared

dl.clear_results()

add the stack object to the stack

dl.add_stack(stack)

run all the transforms and models and return the results in a datframe

results = dl.process_stack()

use this in a notebook to print the results sorted by the highest r2 score

results.sort_values(['data', 'r2_score'], ascending=False)