)
# Dimensionality reduction using MCA:
# Applies only for categoric columns
cols_by_type = sysarmy_analysis.group_cols_by_type()
cols_categoric = sysarmy_analysis.get_cols_by_type(cols_by_type, ["object"])
sysarmy_analysis.reduction_dims(
cols_categoric,
method="mca",
final_number_dims=2,
visualize=True
)
sysarmy_analysis.clusterization(
cols_to_standard,
method="dbscan",
visualize=True
)
sysarmy_analysis.dummy_cols_from_text(col="tecnologies", sep=",", n_cols=15)
print(sysarmy_analysis)
# Salary prediction with linear regression with cleaned columns, no dim reduction
sysarmy_analysis.linear_regression(
col_to_predict="sueldo_mensual_bruto_ars",
cols_to_remove=["PC1", "PC2", "MC1", "MC2"],
graph=True,
)
stackoverflow_analysis.replace_missing(all_cols, method='remove')
stackoverflow_analysis.replace_outliers(cols_numeric, method='drop_iqr')
# stackoverflow_analysis.describe(graph=True)
# ----------------------------------------------------------------------------------
# Data processing
all_cols_to_standard = cols_numeric
stackoverflow_analysis.standardize(all_cols_to_standard, 'z_score')
# Dimensionality reduction using PCA:
# Applies only for numeric columns, requires standardized values
stackoverflow_analysis.reduction_dims(all_cols_to_standard,
method='pca',
final_number_dims=2,
visualize=True)
stackoverflow_analysis.clusterization(all_cols_to_standard,
method='dbscan',
visualize=True)
# stackoverflow_analysis.dummy_cols_from_text(col='technologies', sep=',', n_cols=15)
# print(stackoverflow_analysis)
# ----------------------------------------------------------------------------------
# stackoverflow_analysis.reset()
# print(stackoverflow_analysis)
# stackoverflow_analysis.save(output_path / 'stackoverflow_survey_analysed.csv')
# print(stackoverflow_analysis)