Datacuration

Objective:

  • Understand the whole process of machine learning for Fraud Detection in Barclays
  • Understand everything there is to data involved in the workflow
  • Develop frameworks and tools for automation of data pre-processing and relevant data pipelines
  • Improve the efficiency of the whole process as much as possible

Approach:

  • Learnt about several data sources involved in fraud detection
  • Developed a system for automation processing and reporting on data.
    • Consisted of several modular subcomponents
    • Scalable and time-tested on large datasets
    • Works on both bigdata and smalldata
    • Generated automated reports for different uses and purposes
    • Dashboard for better visualization

Results:

  • Removed manual efforts in several key work pieces
  • Improved completion timeline from months to hours
  • Automated workflows for better maintenance