Streamlining Product Extraction and HS Code Mapping
Our client, a global consulting firm, faced challenges in developing an accurate model for their public sector client. They enlisted the help of our Data Analytics team to transform a substantial dataset of 200,000 purchase records from various sources. The goal was to extract the purchased products and perform a precise and efficient assessment of HS Codes based on country classifications.
The team began by reducing the dataset, achieving a 40% impact through data cleaning and preprocessing, filtering irrelevant information, and eliminating duplicates. They clustered and removed certain service activities and medical products. They then used natural language processing to convert activity descriptions into concise product names and mapped HS Codes to products with OpenAI’s GPT-4, ensuring accurate mapping.
By leveraging the combined power of AI and data analytics models, the client achieved high accuracy and efficiency in product classification, significantly reducing manual effort and time. In conclusion, the collaboration with the global consulting firm led to a streamlined process for product extraction and HS Code mapping, highlighting the effectiveness of AI-driven solutions in addressing complex data challenges.
Client Challenge
The client was struggling to build an accurate model for their end client in Public Sector. Reduction of the initial dataset size The needed support was regarding the transformation of large a dataset comprising 200,000 lines of purchases from various counterparts, to extract acquired products and conduct a precise and efficient assessment of HS Codes at the accurate level of the country classification.
Our Solution
- Dataset Reduction Impact of 40%
-Using Data cleaning and preprocessing methods as the use of “stop word” techniques to filter out irrelevant information and remove duplicates.
–Clustering of service activities, medical products that must be removed from the dataset.
- Automated Data Extraction of products
-Leveraging natural language processing to transform detailed activities descriptions into clear, concise product names.
- Efficient mapping of HS Codes by product
-Utilizing the advanced capabilities of OpenAI’s GPT-4 to analyze and extract key information from product descriptions to be able to map the product to the corresponding HS Code.
Outcome
By coupling the power of AI and Data Analytics models, the client got a high accuracy and efficiency in product classification, significantly reducing manual effort and time.
Other Client Stories
View All
