Skip to main content

Data Analytics, Data Sources, and Architecture for Upgraded European Customs Risk Management

Periodic Reporting for period 2 - PROFILE (Data Analytics, Data Sources, and Architecture for Upgraded European Customs Risk Management)

Reporting period: 2020-02-01 to 2022-02-28

Ensuring security and countering fraud in international trade are key priorities for EU customs administrations, under the heightened risk of transnational crime and terrorism. At the same time, fast and reliable trade flows are critical to the growth and competitiveness of the EU economy. With growing trade volumes, tight resource constraints, and increasing demands for facilitated cross-border traffic, the EU customs administrations are looking for new ways to maintain regulatory control without disrupting trade and commerce. To find the balance between regulatory control and trade facilitation, risk-based and data-driven approaches to customs operations have formed the foundation of customs control activities in Europe since many years. Risk-based controls allow customs to focus on high-risk traffic, facilitate low-risk traffic, and this way oversee cross-border trade without disrupting the flow of goods. However, the risk-based approach also requires EU customs to expand their capability to collect reliable data on trade flows and to analyse this data for effective cargo targeting.

To address this challenge, PROFILE developed and tested modern data analytics tools and explored the value of external data sources for the benefit of customs risk management in the EU. This work was guided by six overall project objectives:
1. Leverage the use of the Advance Cargo Information with improved data access and state-of-the-art data analytics
2. Lower risks of cross-border crime and terrorism, especially CBRNe threats, fiscal fraud, drug trafficking, and counterfeiting
3. Overcome capacity shortages across EU member states
4. Enhance information sharing among European customs and with security, border control, and law enforcement agencies, especially at the EU external border
5. Complement national customs risk management programs
6. Design and prototype a pan-European Risk Data Sharing Architecture
PROFILE results can be grouped under the three PROFILE legacy pillars:
1. Data analytics methods and techniques & a variety of data sources
2. Data sharing, architectures & semantic technologies.
3. Evaluation frameworks & organizational aspects & capacity building

Results in the first Pillar include progressive versions of a customs information portal that helps customs to compare the declared value of imported e-commerce goods against prices of similar goods sold online. These innovations also encompass techniques for exploiting data from e-Commerce platforms for customs risk assessment purposes. PROFILE has also developed a process for using documents and data from TradeLens in the context of a Dutch-based tyre importer. PROFILE partners also developed an interactive user interface for visual data analytics purposes; first version of a deep learning model with supervised and unsupervised learning; first version of an autoencoder model for the detection of outliers and anomalies in new declarations; and, a variety of analytical methods applied for the prediction of accuracy of HS codes, techniques including Natural Language Processing and Random Forest.

The second pillar is about data sharing, architectures & semantic technologies. The Pillar II project results include Semantic Data Modelling and Graph Analytics for Customs Risks; Data linking between customs data sources and external data; New and improved practices for cleansing data for customs risk assessment purposes; A new risk model and process for enriching ENS data from various Member States with external commercial data on operators; A process for comparing exports and imports between an EU and non-EU country; and, Novel machine-learning pre-processing procedure tailored to customs data.

The third Pillar is about “Evaluation frameworks & organizational aspects & capacity building.” PROFLE developed two evaluation frameworks that capture key concepts, requirements and evaluation approaches on how customs can make better use of data and data analytics. The first one is “Evaluation framework for data analytics in risk management”, which summarizes the key experiences that customs partners have had in PROFILE. “PROFILE data analytics value analysis framework” provides insights for customs that seek to incorporate data analytics into their risk management processes and to evaluate the value that data analytics could bring to their operations.

Altogether, PROFILE work has opened many avenues for further exploitation outside the immediate project context. Each PROFILE partner has charted their individual exploitation paths, with details about further R&D, operationalisation, and commercialisation activities. The five PROFILE customs partners plan to exploit technical and organizational innovations of PROFILE mainly to improve their operational customs risk management processes and performance. PROFILE consortium also includes two partners that have commercial interests in exploiting the project’s technologies and other outcomes through commercialisation activities. Lastly, the seven research-oriented PROFILE partners see the project outcomes primarily as a way to build training programs, improve consulting services and/or to conduct further research.
The PROFILE results contribute to the development and uptake of modern data analytics techniques — such as machine learning, natural language processing, and graph-based analytics techniques — in the customs risk management context. PROFILE innovations comprise technical algorithms and software, visualization tools, strategies for cooperation, techniques for data management, and conceptual frameworks, all tested in practice with customs end-users. These solutions were found to help customs administrations to collect and organize information, mine and visualize large datasets, and to make better use of control feedback and inspection outcomes. PROFILE also assessed the value of several pre-existing and new data sources for customs risk management. For example, PROFILE acquired access to external data sources that contained information about EU-bound containers (INTTRA), container ships (eeSea) and about traders & importers (Dun&Bradstreet and Orbis). Besides concrete innovations, PROFILE also discovered valuable lessons on approaches that do not easily work for Customs that wish to become more data-driven organizations. Critical success factors include close cooperation between Customs and technical experts, meticulous data preparation, and proactive tackling of any issues hindering the joint efforts by customs, technical partners, and suppliers of external data.

In the broader societal context, PROFILE results contribute to supply chain security and the fight against cross-border trafficking. PROFILE innovations in data and data analytics help customs to leverage digital information for data-driven risk-based border controls, leading to higher situational awareness about evolving security threats. This information helps customs to make timely decisions about the optimal use of limited inspection resources (staff, expertise and technology) and this way detect a higher percentage of illegal goods (e.g. drugs, weapons, fiscal contraband and explosives). PROFILE contributions to data-driven, risk-based and selective customs controls also are expected to result to faster and more predictable cross-border trade and logistics operations. This trade facilitation benefit strengthens the competitiveness and viability of EU-based industries.
Logo