Embedded Dashboard Generator for EasyEMI - EDGE Qorus Banking Innovation Awards 2023

The consumer finance team of HDFC Bank caters to 2M potential customers every month with average monthly business of around $228M. This volume is spread across 11 different categories - Electronics - Mobiles and Wearables - E-tailing - Apparels - Travel - Home Decor - Luxury - Healthcare - Insurance - Education - Others To keep a track of the business achieved under each of these categories, the team manually gathers more than 12 files from various aggregators and creates a one view. The business challenges faced were: Data in silos: Data is stored and processed in different formats because they are sourced from different applications that don't communicate with each other. Time-consuming data analysis: Manually consolidating data from different sources and analysing it is extremely time-consuming and error prone. Limited data visibility: When data is stored in different systems, it can be difficult to get a comprehensive view of business performance. Lack of collaboration: When data is stored in different systems, it can be difficult for different teams to collaborate effectively. To solve these issues, we developed a web application (EDGE) which could be at the fingertips of the stakeholders, so that they could download a one view of Consumer Finance business anytime. EDGE consolidated data form different silos and created near real time dashboards that not only provided immediate insights into business performance, but also provided the stakeholders with more reliable data reporting with help of extensive text mining which eliminated any scope of human error. A total of 20 files were received every week which included 12 Raw Files (Loan boarding Files and 8 Master Files for Segment Mapping. The full process of Segment Dashboard for Consumer Durables using Embedded Analytics can be classified into 3 parts 1. Data Understanding Multiple challenges were discovered while understanding the vast quantum of data coming every week a.) Multiple Data Sources – Since the 12 raw files incoming every week had different sources, the data structure needed to be consistent to stitch them together to create a single report b.) Inconsistency in data structure – Since collection of data is still a manual process, there were a lot of inconsistencies across the framework of files - File Name, Column Names and Sheet Names across the weeks. c). Multiple sources for Segment mapping- Finally, after data is consistent across time and format, the segments needed to be mapped for each of the transaction to create a summary. The logic for segment mapping for each of the file varied significantly. 2. Creation of Dashboard After the nascent step of data collaboration from 20 different files, next step is to create the summary in the back end which not only emulates the procedure done manually but also stands as an improvement in data classifications. It was achieved by performing deep text mining on the Manufacturer Name, Merchant Name and the product Category to gap the bridge of any manual misses. 3. Automation Finally, the whole process converged with a simple yet elegant automation pipeline -Raw data files are placed in a pre-determined folder or directory whenever the dashboard is to be generated - When EDGE is triggered, it fetches this data for processing through the back-end with help of python scripts - The transformed data is then used to generate a summary report or dashboard within 15 minutes. The automated system applies specific algorithms, calculations, and formatting to convert the processed data into a dashboard with key metrics and summaries 4. Impact EDGE has had a measurable impact on the future of automation in our organisation Manual tasks: The manual process of making this dashboard took 12–15-man hours every week by 3 different people. EDGE has reduced that time to just 15 minutes. It addressed the challenge of performing manual tasks related to data entry and analysis, which can be time-consuming, error-prone, and resource-intensive. Efficiency: By automating manual tasks, the project helped improve efficiency and productivity, allowing the organisation to focus more time and attention on other important areas of the business. Accuracy: Not only has it reduced the tremendous manual effort which goes into creating this dashboard, it also improved the classification rate of the report by 10-12% each week Innovative: It provides end to end Automation which includes data crunching, data cleaning, text mining, consolidation, segment mapping, LFR/RLFR mapping and summarizing EDGE has directly and indirectly waved the flag for innovation and automation as it had led to increased efficiency, timely results, better decision making and improved outcomes across the board. Overall, EDGE has helped address several key business challenges related to data analysis and management, and provided a scalable and sustainable solution to these challenges