Transforming Data Parsing for BluePallet’s Precise Knowledge Graph Construction

In today’s business landscape, managing extensive unstructured data presents various challenges. Extracting insights from sources like text documents, log files, or web pages is hindered by the lack of a predefined format, introducing complexities in parsing. Complicated entity extraction, variations in formats, and noisy data further complicate the process, demanding robust techniques for accuracy.

Amidst this complex landscape, BluePallet, a pioneering force in the chemical commerce sector, encountered a substantial challenge in efficiently organizing Safety Data Sheets (SDS) to create a detailed knowledge graph. The company’s collaboration with Wisecube proved helpful in addressing these challenges, marking a strategic venture where advanced machine-learning solutions revolutionized data parsing strategies.

This case study explores the strategic collaboration between BluePallet and Wisecube, offering insights into how Wisecube helped BluePallet overcome the complexities of parsing unstructured data and reshaping the dynamics of the chemical industry.

About BluePallet

BluePallet is a virtual marketplace that connects manufacturers with the chemical sector. It offers a comprehensive industrial commerce platform that provides innovative solutions for various aspects of the industry, such as search, logistics, and transactions. Through its groundbreaking approach, BluePallet has transformed the international chemical commerce environment, making it more accessible and efficient for all stakeholders involved.

BluePallet’s platform offers unique features to help users in their sourcing and procurement efforts. By leveraging its network functionality, BluePallet makes it easy to create new sourcing relationships while improving existing ones to ensure optimal results. Users can easily acquire needed products through real-time pricing info, direct communication with sellers, and streamlined procurement processes. 

Objective: AI-Driven SDS Parsing For Actionable Insights

BluePallet’s unwavering dedication to advancing growth, efficiency, and success in the industrial commerce domain led to a strategic partnership with Wisecube. Together, they spearheaded pioneering data parsing strategies, aiming to develop an AI-driven solution. This collaboration seeks to revolutionize the interpretation of Safety Data Sheets (SDS) and other chemical documentation, converting static data into dynamic, accessible, and actionable insights. This initiative aligns with BluePallet’s core mission to transform and optimize the chemical commerce experience through cutting-edge technology and innovation.

The Challenge: Parsing Safety Data Sheets 

Safety Data Sheets (SDS) are vital pillars in BluePallet’s chemical logistics process, offering indispensable insights into the properties, hazards, and safe handling procedures for various chemicals. These documents are ripe with critical safety and product knowledge entangled within dense technical terminologies and varying formats across different manufacturers. For BluePallet, leveraging SDS is fundamental to ensuring the safety, compliance, and informed decision-making across their chemical commerce platform.

A Sample SDS – redacted 

Despite their indispensable nature, Safety Data Sheets (SDS) presented challenges due to diverse layouts, making extraction complex. The inherent complexities of SDS sheets demanded thorough efforts in understanding and interpretation.

Navigating through this complex SDS landscape, BluePallet faced a daunting challenge of efficiently parsing the safety sheets to construct an accurate knowledge graph. Here are some of the hurdles encountered by BluePallet in this process:

  • Complicated Entity Extraction: Disentangling complex entities buried within the dense data of SDS sheets demanded intricate algorithmic parsing.
  • Noisy Data Resulting in Inaccurate Parsing: Overcoming inaccuracies caused by noise in the data became pivotal for precise information extraction.
  • Interpretation Challenges due to Lack of Context: Deciphering information without contextual cues posed challenges in accurately understanding the data’s implications.
  • Interpreting Safety Guidelines: Deciphering and accurately interpreting safety guidelines embedded within the tabular data emerged as a critical challenge, necessitating a nuanced approach to data extraction.
  • Adapting to Diverse Data Formats and Structures: Flexibility in parsing to accommodate varying document structures and data formats proved essential for comprehensive extraction.
  • Requiring High Precision Extraction for Accuracy and Safety: Precision-driven extraction methodologies were crucial to ensure data accuracy and maintain safety standards within the chemical industry.
  • Time-Sensitive Parsing: The urgency to parse SDS sheets efficiently added an additional layer of complexity, demanding a solution that could balance speed and accuracy.

In order to overcome parsing obstacles and create a more robust knowledge graph, BluePallet collaborated strategically with Wisecube.

Wisecube’s Solution: SmartParser Reshapes BluePallet’s Data Landscape

In response to BluePallet’s complex parsing challenges, Wisecube emerged as the catalyst for transformation, offering a comprehensive solution to tackle the complexities of SDS sheet extraction. Utilizing their expertise, Wisecube introduced a versatile approach tailored to meet the unique requirements of BluePallet’s knowledge graph construction.

Wisecube’s AI-Driven Smart Parsing Approach:

Utilizing the NLP Lab delivered by John Snow Labs, Wisecube executed a comprehensive smart parsing strategy:

  • Data Annotation: Before parsing, Wisecube employed the NLP Lab to label and categorize raw data effectively, ensuring the training of models to recognize patterns and structures within SDS.
  • Entity Recognition and Extraction: Trained models focused on recognizing and extracting crucial details such as chemical names, properties, and hazards for comprehensive information capture. These models were intricately designed to produce well-organized information in JSON format, making it readily available for constructing a chemical knowledge graph. 
  • Noise Handling and Data Cleansing: Sophisticated algorithms filtered out errors and irrelevant data, ensuring accurate parsing and relevant information extraction.
  • Contextual Understanding: Wisecube’s models were designed to grasp contextual nuances embedded within SDS, ensuring the extracted data was both accurate and meaningful.
  • Parsing Efficiency: Prioritizing efficiency, Wisecube ensured rapid processing of vast volumes of data sheets without compromising accuracy.
  • Confidence Computation: After parsing, the system calculated a confidence score for extracted data, quantifying the reliability of the parsed information, providing users with insights into data reliability.
  • Human in the Loop Validation: Dual-layered validation involved human experts reviewing and validating extracted data. This process guaranteed that the extracted data aligned precisely with BluePallet’s distinct requirements, meeting the rigorous standards of accuracy crucial for constructing a dependable knowledge graph.

An illustration of Wisecube’s SmartParser solution for BluePallet


Wisecube’s Smart Parser seamlessly integrated into Blue Pallet’s system via a GraphQL Java API and AWS MQ, ensuring robust communication and data exchange. Rigorous testing and refinements prepared the parser to adeptly handle the intricate nature of chemical data sheets. The parser combined the following key components for precise and reliable parsing of unstructured data:

SmartParser and Amazon MQ architecture for message handling

  • Wisecube API: This gateway enables users to effortlessly upload, track parsing status, validate results, and retrieve structured content from PDFs. The API simplifies SmartParser integration into existing workflows.
  • Parser Engine: At the core, SmartParser employs advanced algorithms leveraging ML and NLP for automatic extraction and interpretation of data elements from PDF documents.
  • Parser Verification: Using pre-annotated data on Amazon MQ, SmartParser verifies accuracy by comparing parsed output against annotated data, ensuring reliability.
  • John Snow Labs (JSL) Annotator: Incorporating JSL’s PDF annotation tool enables contextual feedback and correction of misidentified values by the parser. Correctly identified values by the annotator refine the parser model for accurate parsing.

With the implementation of Wisecube’s SmartParser, BluePallet gained the capability to extract valuable insights from SDS sheets. The parsed data not only proved reliable and well-organized but also enriched with essential details, contributing to the construction of a highly accurate knowledge graph.

Outcomes of Wisecube’s Intervention

The implementation of Wisecube’s SmartParser proved instrumental in empowering BluePallet to overcome the challenges associated with parsing complex SDS sheets. The outcomes were transformative:

  • Reliable Data Extraction: Wisecube’s SmartParser enabled BluePallet to extract valuable information from SDS sheets with reliability, ensuring the accuracy and integrity of the parsed data.
  • Optimized Workflow Efficiency: Wisecube enabled Bluepallet to streamline the handling of extensive document volumes, significantly improving workflow efficiency.
  • Safety Compliance and Hazard Reduction: Through meticulous data interpretation, Wisecube helped BluePallet mitigate potential safety risks and ensure compliance with safety regulations.
  • Structured Data Output: The ML model generated well-structured data in a JSON format, providing an organized foundation for the construction of BluePallet’s knowledge graph.
  • Dynamic Data Utilization: Wisecube’s solution empowered BluePallet to convert static data sources into dynamic, actionable insights, revolutionizing their approach to information utilization.
  • Human-Verified Precision: The integration of HITL validation ensured precision and accuracy, aligning the extracted data precisely with BluePallet’s requirements.
  • Elevated Safety Standards: The improvement in confidence and safety in handling chemical products benefited both workers and end-users alike.
  • Environmental Responsibility: BluePallet advocated for responsible chemical transportation and usage, fostering environmental sustainability through conscientious practices.

With a precise and reliable knowledge graph at its core, BluePallet is now set to revolutionize the chemical commerce sector, facilitating seamless interactions and driving innovation within the global chemical community.


The collaboration between Wisecube and BluePallet surpassed parsing challenges, resulting in improved understanding of chemical safety guidelines. The synergy of advanced technology and human validation set BluePallet into a new era of efficiency, ensuring reliable data extraction and structured outputs. This cooperative success positions BluePallet as an industry innovator, ready to redefine the chemical commerce landscape with an enriched and accurate knowledge foundation.

​This case study not only serves as a success story but also highlights the transformative impact of integrating Wisecube’s SmartParser. It sets a precedent for industries, showcasing how the combination of machine learning and human validation drives efficiency, accuracy, and robust knowledge construction. BluePallet’s experience underscores innovation in industrial commerce, inspiring businesses in data-intensive sectors.

For an in-depth exploration of Wisecube’s innovative data parsing strategy with BluePallet, tune into our NLP Summit Webinar here.

Table of Contents

Scroll to Top