🚀 Quickstart

This guide shows how to quickly generate an EDA report using the owlmix package.

1. Install dependencies (if not already installed):

   pip install -r requirements.txt
   

2. Generate a report and render HTML

   import warnings
   from pathlib import Path
   
   # Suppress user warnings for cleaner output
   warnings.simplefilter('ignore', category=UserWarning)
   
   # Import necessary modules from owlmix
   from owlmix.utils.sample_data_generator import create_sample_data
   from owlmix.reporting import ReportBuilder, ReportHTMLRenderer
   
   # Set up paths
   CURRENT_DIR = Path(__file__).parent
   OUTPUT_DIR = CURRENT_DIR / "output"
   OUTPUT_DIR.mkdir(exist_ok=True)
   
   # Generate sample data (you should replace this with your own dataset)
   df = create_sample_data(n=500)
   
   # Build the report
   report_builder = ReportBuilder(
       df=df,
       target_col="sales",
       date_col="time"
   )
   
   # Update configuration for ACF/PACF and VIF
   report_builder.config.update_config(
       acf_pacf={
           "columns": ["sales", "tv_spend"],
           "n_lags": 5,
           "precision": 3
       },
       vif={
           "precision": 2
       },
       correlation={
         # "columns": ["sales", "tv_spend"],
         "n_lags": 8,
         "precision": 5
     }
   )
   
   # Add all sections to the report
   report_builder.add_all_sections(verbose=True)
   
   # Build and save the report as JSON
   report = report_builder.build()
   report_builder.save(OUTPUT_DIR / "result.json")
   
   # Render HTML from the JSON report
   renderer = ReportHTMLRenderer()
   html_str = renderer.render_from_json(OUTPUT_DIR / "result.json")
   renderer.save_html(html_str, OUTPUT_DIR / "report.html")
   

3. View the generated HTML report Open output/report.html in your browser to see the EDA report.

Load Config from Dict, YAML, or JSON

You can also load section configuration from a Python dictionary, YAML file, or JSON file.

# 1) From Python dict
config_dict = {
    "acf_pacf": {
        "columns": ["sales", "tv_spend"],
        "n_lags": 10,
        "precision": 2,
    },
    "ccf": {
        "feature_columns": ["tv_spend", "digital_spend"],
        "max_lag": 3,
    },
}
report_builder.config.update_config_from_dict(config_dict, strict=True)

# 2) From YAML file
report_builder.config.update_config_from_yaml("report_config.yaml", strict=True)

# 3) From JSON file
report_builder.config.update_config_from_json("report_config.json", strict=True)

Notes:

• strict=True raises an error for unknown section keys.

• The top-level config must be a dictionary.

• Each section value must also be a dictionary.

Back to Home