BDIViz is a powerful, interactive tool designed as an extension to BDIKit to assist biomedical researchers and domain experts in performing schema matching tasks. Built to address the challenges of matching complex biomedical datasets, BDIViz leverages a visual approach to streamline the process and enhance both speed and accuracy.
Key features of BDIViz include:
- Interactive Heatmap for exploring and comparing matching candidates.
- Value Comparisons Panel for analyzing similarities between attributes.
- Detailed Analysis Panel offering in-depth insights into attribute value distributions.
- Filtering & Refinement Tools to customize and adjust matching candidates based on datatype and similarity scores.
- Expert-in-the-Loop Workflow allowing users to iteratively accept, reject, or refine matches, keeping the expert in control of decision-making.
BDIViz is designed to be integrated with Python notebooks, providing a flexible and easy-to-use tool for domain-specific schema matching in biomedical research and beyond.
To use BDI-Viz, install it using pip:
pip install bdi-vizBDI-Viz 1.0 is built leveraging Panel. The application is designed to provide a user-friendly interface on jupyter notebooks. Where users can explore the schema matching recommandations, interact with the result, and pass them to the next step of the data integration process.
import pandas as pd
from bdiviz import BDISchemaMatchingHeatMap
# Load the data
source_df = pd.read_csv('data/source.csv')
target_df = pd.read_csv('data/target.csv')
# Render the BDI-Viz Heatmap
heatmap_manager = BDISchemaMatchingHeatMap(
source=source_df,
target=target_df,
top_k=20,
)
heatmap_manager.plot_heatmap()The following interface will be displayed in the jupyter notebook:
For more information, please refer to the documentation.