How to set up an analytics platform leveraging Microsoft technologies for an NBFC?
Typically, a non-banking financial services company will have a very small internal IT team and will be dealing with a lot of data management residing in their loan origination systems, loan management systems pertaining to all the different customer segments they focus upon.
These systems and data are subject to change and when there is a change, there are challenges with the data migration. Hence it makes strategic focus for NBFCs to look for a common data mart that can integrate from different venues, a data mart that can outlast the changes and updates.
There are numerous paths that can be taken. A lot of LMS products offer an analytics platform in built. They come with a cost and customization that might or might not suit the buyer. There are also isolated products where the data is extracted and pushed onto the platform. This option might be time consuming. Looking at these choices might be overwhelming for an NBFC starting out or looking to change or update their current set up.
Fortunately, the array of cloud-based solutions and visualization tools made available by Microsoft, can aid in setting up an agile and cost-effective analytics platform which will prove decisive in the growth of NBFC organizations.
Below a road map to set up an analytics platform leveraging the Microsoft technology stack:
Starting from the understanding why there is a need for a data set up, the identification of inadequacy in all stages of business processes starting from application of loans, disbursal, collection, and further analytics or lack thereof. It also involves vetting of previously existing
- Data acquisition systems be it mobile apps, or any other high-tech LOS/LMS or even a plain vanilla excel sheet
- Data warehouses that can be NoSQL or SQL databases or just a folder on your desktop
- Analytics platform if present, the math and the technology involved.
With a comprehensive view of what our requisites and predicaments are, we can decide what needs to be done.
During sales or following up on leads, entering the application details and while collecting the payment. Current process can involve entering the data into a SaaS or excel sheets. The less accessible and vexing the medium in which the data is entered, the more it contributes to aberrations in the data and slows down the process. Hence the practice should be made as convenient as possible.
There are cost effective options available in the market, but all these options involve a huge effort, cost and time in building such apps. Microsoft now offers Power Apps that lets anybody build an app without the understanding of complex coding languages. It can be deployed swiftly and scaled up.
Extraction transformation and loading might not be a part of the equation if the major source of data entry is in excel sheets. But once the data size increases, ETL very much becomes a part of it. The data must be collected from varied sources, validated, transformed, and passed into a desirable storage facility. The routine must be quick irrespective of the size of the data. ETL can be done using automation accounts on Azure with python or power-shell scripts depending on the size of the data. ETL on a large scale is easily possible leveraging Azure Data Factory.
Storage becomes the foundational block for the operations and analytics of an NBFC data. The preceding storage system might be outdated, slow, or just inefficient in nature. The storage facility must be rapid in storing, retrieving and other internal calculations. This can lead to great improvement in the efficiency of the operations.
The loan tracking system, repayment schedules, overdue calculations are all directly dependent on the quality of the storage. The choices in such storage is abundant now. An Azure Cloud account let us run storage accounts, supports automation that gives us ETL capability and myriad of other features that are handy. Storage can start from a simple cloud storage account to varied options like Azure SQL, Azure Cosmo DB, NoSQL database and much more.
For exploratory analytics in a small scale, Microsoft Excel can be enough. But once the size of the data becomes enormous, it is natural to transcend to a tool designed for such purposes. Data viz tools make it easier to create visuals for the data collected from variety of sources and giving access to multiple charts for the data to fit in, paving the way to gain insights and make better business decision. Power Bi the data visualization platform offered by Microsoft seamlessly integrates to excel and the azure cloud. It enables self-service business intelligence capabilities for business users
Machine learning would be a big leap from the analytics provided from the statistical methods and Exploratory data analytics. Machine learning has multiple use cases in NBFC, starting from credit risk modelling to NLP on customer feedback. This setup can prove convenient for starting from basic operational dashboards on Disbursements, Collections and PAR reports, to slicing and dicing of data for descriptive and/or diagnostic analysis, to an elaborate machine learning model for credit risk modelling.
Azure cloud-based components help us produce an extremely agile and customized analytics solution that can be built leveraging mobile, database capabilities within a month.
Blog disclaimer:
This is a professional weblog, and we have invited experts to share their thoughts, expertise , perspectives and knowledge. The opinions expressed here are purely representing their personal views and not those of any institution, employer or company.
