A Novel Visual Analysis Method of Food Safety Risk Traceability Based on Blockchain

Zhihao Hao, Dianhui Mao, Bob Zhang, Min Zuo, Zhihua ZhaoInternational Journal of Environmental Research and Public Health 2020 · DOI

TL;DR

This paper introduces a novel approach to food safety by combining blockchain for secure data storage with advanced visualization techniques for intuitive risk analysis. By leveraging Hyperledger Fabric and visual tools like heat maps and force-directed graphs, the system aims to provide regulators with a clear understanding of food safety risks and their origins, enabling more effective regulatory strategies.

What Problem Does This Paper Address?

Current food traceability systems often suffer from data tampering and a lack of intuitive methods for analyzing the root causes of safety risks. This makes it difficult to quickly recall unsafe products and for regulatory bodies to effectively manage food safety. Existing systems struggle with incomplete information storage, the ease with which data can be forged, and the complexity of analyzing vast amounts of food safety data. This complexity hinders timely decision-making and the formulation of targeted regulatory interventions.

Methods

The proposed solution integrates two key technologies:

  • Blockchain for Data Storage: Hyperledger Fabric is used to build a distributed and tamper-resistant platform for storing food sampling data. This ensures the authenticity and integrity of the information. A custom data structure model is designed to accommodate different user roles and data types.
  • Visualization for Risk Analysis:
    • Quantitative Risk Assessment: A model is established to quantitatively analyze food safety risks based on the failure rate and deviation of unqualified detection data.
  • Macro Analysis (Heat Maps): Heat maps are employed to visualize the geographical distribution of food safety risks, highlighting "hotspots" where unqualified products are frequently detected. This involves grid clustering, Gaussian fitting, and color mapping to represent risk intensity.
  • Micro Analysis (Migration Maps & Force-Directed Graphs): To understand the origins and flow of risks, migration maps and force-directed graphs are used. Migration maps show general flow directions, while force-directed graphs illustrate specific pathways and relationships between locations, enabling detailed traceability analysis of unqualified products.

Datasets

The primary dataset used for evaluation is the Food Sampling Data for the full year of 2016. This dataset comprises 341 types of food and over a million items. For their case study, the authors focused on 19,578 aquatic products, including fish, shrimps, crabs, shells, and mollusks. To protect merchant privacy, address hiding and other processing were applied to the data.

Key Insights and Ideas

  • Blockchain as a Trust Anchor: The inherent properties of blockchain (distribution, tamper-resistance) are crucial for establishing trust in food safety data, overcoming the limitations of traditional centralized systems.
  • Bridging Data Complexity and Human Intuition: The core innovation lies in translating complex, quantitative food safety data into easily understandable visual representations. This allows for rapid identification of high-risk areas and the tracing of product origins.
  • Holistic Risk Management: The approach moves beyond simply identifying risks to understanding their causes and origins through detailed traceability, enabling more targeted and effective regulatory interventions.
  • Role-Based Access Control: The system utilizes different user roles (manufacturer, processor, inspector, transporter, distributor) with varying permissions, ensuring data security and appropriate access for different stakeholders.

Key Results

The experimental results demonstrate the effectiveness of the proposed method:

  • Visualized Risk Distribution: Heat maps successfully illustrated the geographical distribution of food safety incidents over time, showing a trend of incidents emerging in inland areas and concentrating in specific regions.
  • Traceability of Unqualified Products: Migration maps and force-directed graphs were able to trace unqualified aquatic products back to their origins from multiple locations, providing concrete evidence for regulatory action. For instance, a specific unqualified product (a₁ in Grid A) was traced back to eight different source locations.
  • Foundation for Regulatory Strategy: The visual analysis provides a clear basis for regulators to identify high-risk areas and understand the flow of unqualified products, facilitating the formulation of stricter regulations, increased spot checks, and enhanced penalties.

Limitations and Future Work

The authors acknowledge several limitations:

  • Blockchain Scalability and Speed: Blockchain technology, particularly Hyperledger Fabric, faces challenges related to the speed and scalability of block generation, which can impact data processing efficiency.
  • Data Quality Dependence: The accuracy and validity of the visual analysis are heavily dependent on the quality of the input sampling data. Sampling errors can significantly affect the results.

Future work should focus on:

  • Developing effective methods to address the speed and scalability issues of blockchain.
  • Implementing strategies to mitigate the impact of sampling errors and improve data quality.
  • Further refining visualization techniques for even more accurate and insightful analysis.

Summary updated

Abstract

Current food traceability systems have a number of problems, such as data being easily tampered with and a lack of effective methods to intuitively analyze the causes of risks. Therefore, a novel method has been proposed that combines blockchain technology with visualization technology, which uses Hyperledger to build an information storage platform. Features such as distribution and tamper-resistance can guarantee the authenticity and validity of data. A data structure model is designed to implement the data storage of the blockchain. The food safety risks of unqualified detection data can be quantitatively analyzed, and a food safety risk assessment model is established according to failure rate and qualification deviation. Risk analysis used visual techniques, such as heat maps, to show the areas where unqualified products appeared, with a migration map and a force-directed graph used to trace these products. Moreover, the food sampling data were used as the experimental data set to test the validity of the method. Instead of difficult-to-understand and highly specialized food data sets, such as elements in food, food sampling data for the entire year of 2016 was used to analyze the risks of food incidents. A case study using aquatic products as an example was explored, where the results showed the risks intuitively. Furthermore, by analyzing the reasons and traceability processes effectively, it can be proven that the proposed method provides a basis to formulate a regulatory strategy for regions with risks.
BibTeX
@article{hao2020novel,
  author = {Zhihao Hao and Dianhui Mao and Bob Zhang and Min Zuo and Zhihua Zhao},
  title = {{A Novel Visual Analysis Method of Food Safety Risk Traceability Based on Blockchain}},
  journal = {International Journal of Environmental Research and Public Health},
  year = {2020},
  doi = {10.3390/ijerph17072300},
}
Figure