Case Studies
Live dashboards on Tableau
Problem Statement
Client wanted to enable constant monitoring of KPIs by multiple dimensions such as client, dept size, portfolio, demographics, status etc. Realtime consolidation across data sources was further complicated by manual processing that made periodic refreshes a time kill and frustrating exercise.
Solution
Internal data processes were explored in depth to understand the right time and place/system to extract the required information. An elegant solution involving application of Tableau connectors, SQL and R scripts was created to reach the desired goal.

Excel to text file conversion in a pre-defined format
Problem Statement
Solution

Payment scoring algorithm creation
Problem Statement
Client was aiming to optimize their collection strategy by prioritizing on accounts that are more likely to pay. Attempting the above basis just some high-level borrower attributes wasn’t
giving reliable results.
Solution
All available data sources including Borrower data, payment history, user demographics, account attributes, existing call data etc. were collated together. The consolidated dataset was then used to create a robust logistic regression model that was further validated to ensure efficacy of results.

Automated data extraction to generate excel reports
Problem Statement
Available reports from tools such as Qlaw or Beam were insufficient to provide the complete required information quickly and in a single place. Extracting more information from the backend tables was complicated & time consuming task and even the available reports required manual pull every time.
Solution
Thorough exploration of the backend data system helped identify all involved pieces required to build the excel reports without missing critical information. Once identified, these data files were set to feed the local SQL database and subsequently the reporting scripts that would create the final output automatically

Extract Transform & Load (ETL) process integration
Problem Statement
Client was struggling to onboard a new dialer system. The key issues involved seamless extraction of data from the CRM system, exclusion of accounts post application of relevant rules and completing the cycle by uploading the result files back to the CRM system.
Solution
Pre-programmed scripts were put in place to pick/drop the information to andfrom the CRM system while working within SFTP folders. Amid the file movement, data cleaning occurred every morning before the upload & the result files (generated every hour) would be transformed to acceptable format before getting pushed to the CRM system
