Standardise variable names in the docs (#338)

* selected_columns -> feature_column_names

* change reference and analysis to reference_df and analysis_df

* Merged with main

---------

Co-authored-by: Niels Nuyttens <niels@nannyml.com>

README.md

@@ -159,9 +159,9 @@ import pandas as pd
 from IPython.display import display
 
 # Load real-world data:
-df_reference, df_analysis, _ = nml.load_us_census_ma_employment_data()
-display(df_reference.head())
-display(df_analysis.head())
+reference_df, analysis_df, _ = nml.load_us_census_ma_employment_data()
+display(reference_df.head())
+display(analysis_df.head())
 
 # Choose a chunker or set a chunk size:
 chunk_size = 5000
@@ -175,8 +175,8 @@ estimator = nml.CBPE(
     metrics=['roc_auc'],
     chunk_size=chunk_size,
 )
-estimator = estimator.fit(df_reference)
-estimated_performance = estimator.estimate(df_analysis)
+estimator = estimator.fit(reference_df)
+estimated_performance = estimator.estimate(analysis_df)
 
 # Show results:
 figure = estimated_performance.plot()
@@ -192,8 +192,8 @@ univariate_calculator = nml.UnivariateDriftCalculator(
     chunk_size=chunk_size
 )
 
-univariate_calculator.fit(df_reference)
-univariate_drift = univariate_calculator.calculate(df_analysis)
+univariate_calculator.fit(reference_df)
+univariate_drift = univariate_calculator.calculate(analysis_df)
 
 # Get features that drift the most with count-based ranker:
 alert_count_ranker = nml.AlertCountRanker()
@@ -214,10 +214,10 @@ figure = univariate_drift.filter(period='analysis', column_names=['RELP','AGEP',
 figure.show()
 
 # Get target data, calculate, plot and compare realized performance with estimated performance:
-_, _, analysis_targets = nml.load_us_census_ma_employment_data()
+_, _, analysis_targets_df = nml.load_us_census_ma_employment_data()
 
-df_analysis_with_targets = pd.concat([df_analysis, analysis_targets], axis=1)
-display(df_analysis_with_targets.head())
+analysis_with_targets_df = pd.concat([analysis_df, analysis_targets_df], axis=1)
+display(analysis_with_targets_df.head())
 
 performance_calculator = nml.PerformanceCalculator(
     problem_type='classification_binary',
@@ -227,8 +227,8 @@ performance_calculator = nml.PerformanceCalculator(
     metrics=['roc_auc'],
     chunk_size=chunk_size)
 
-performance_calculator.fit(df_reference)
-calculated_performance = performance_calculator.calculate(df_analysis_with_targets)
+performance_calculator.fit(reference_df)
+calculated_performance = performance_calculator.calculate(analysis_with_targets_df)
 
 figure = estimated_performance.filter(period='analysis').compare(calculated_performance).plot()
 figure.show()