Case Study-

NLP-based Platform
for a leading business conglomerate

The Client

The client is a Fortune 500 company, a leading conglomerate and one of the biggest private sector corporations. The company extends cutting-edge products and services across energy, petrochemicals, textiles, natural resources, retail and telecommunications.

The Business Situation

The client wanted to setup a system for their HR department that collects all data queries & feedback fed by employees during their training sessions across the organization. They engaged Diaspark to scale up their HR team with an innovative semantic engine capable to search and visualize training programs’ related queries.

DIASPARK’S SOLUTION

Diaspark helped the client with a text mining solution to encode semantics of natural language elements for a simple English language search query. Our development team implemented a configurable NLP pipeline using a prototype of algorithms to return relevant visualized graphical data as output.

Key Functionalities
  • Integration of a NLP pipeline into the system that converts input in normal English language into SQL statement from the schema integrated
  • Implementation of the process of data abstraction using Python & framework Flask for textual analysis including text categorization, phrasing, semantic analysis etc.
  • Conversion of text to graphs using Matplotlib plotting library for comparisons of datasets having different domains
  • Generation of results in form of visual representations such as pie charts, bar graphs etc.
Benefits To The Client
  • Augmented the spectrum of HR operations and compliance with actionable data points
  • Speedy decision making for internal teams with ad-hoc report generation
  • Improved turnaround time, boost efficiencies and discovery experience through automation
  • Provided a 360-degree view of structured and unstructured data by unlocking actionable insights
  • Enhanced knowledge sharing, planning, and exploration of HR operations with digitized search and analytics across data