Shipment entity dashboard
This allowed analysts to proactively pinpoint patterns of
suspicious activity, detect illegal actions and investigate
suspected perpetrators of customs fraud and smuggling,
rather than relying on random inspections.
In addition to the platform’s built-in data fusion, data
visualization and analytics capabilities, NEXYTE’s expert
team of data scientists quickly developed a tailored,
machine learning-based model to automatically score
every incoming package according to its likelihood of
involvement in customs fraud or tax evasion. The customs
organization supplied 2 months' worth of incoming shipment
data, which encompassed over 130 million records from
15 different data sources – including 1.6 million
merchandise records.
Visual link analysis between a suspicious shipment of
merchandise and other customs-related entities
Manifests
Invoices
Incidents
Data records analyzed
Table Records Columns Unique Keys
Merchandise 1.6M 38 1,400,000
Carrier 800K 10 665,357
Manifest 7.7M 7 561,289
Containers 1.3M 7 597,162
Invoices 2.9M 20 1,165,700
Dates 3.3M 7
Identifiers 8.2M 9 1,426,918
Contributions 4.3M 9 1,445,513
Batches 23M 31 1,449,775
Permissions 2.3M 12 351,397
Batch level identifiers 54M 10
Batch level rates 12.3M 10 954,457
Batch level contributions 12.4M 10 954,388
Incidents 230K 21 150,276
Selection 3.1M 14