The customer was a Telecom giant. The client wanted to transform their databases to Azure with an intention to standardize database technology stack and reduce TCO by adopting cloud. The customer was provided with optimum solution to migrate databases connecting to their mission critical applications to Azure VM Infrastructure As Service (Iaas), Azure SQL or Azure SQL Managed Instances as Platform as a Service (PaaS) model. The challenge in this project was to provide a solution which will provide a migration with minimal downtime despite of their large size. The solution included seamless access with High Availability with in SLAs defined by client.

The application tracked points and clicks by customer's website users. The raw data collected was cleansed to remove redundant & inconsistent data. The incomplete data was normalized using master data. The session related parameters were then identified. These data was then subjected to knowledge discovery using data mining techniques. FP-Growth data mining techniques was used to identify association to understand frequent patterns of users of web sites. The web site users were provided with personalized messages and offers by applying k-means technique to cluster website users in various clusters or groups. Offer exclusion &   inclusion were prepared using responses from website and classifying e-mails from customer's clients. All data analyzed was provided using graphical visualization tools.

Analyzed AS-IS database technology landscape in client environment. The requirements & pain points provided by client were analyzed. Since current database technology used had drawbacks in addressing client’s requirements, a framework was developed to evaluate and identify most suitable database or NoSQL technologies. A Proof of Concept was developed on candidate database or NoSQL products to finalize on most suitable technology. Architecture & technology solution was developed using Hadoop Eco System & Terradata Database which were finalized technologies. Hadoop tools such as Hive, Hbase and Map Reduce Framework were used mainly in proposed solution.

The aim of developing customer data services & hub was to have centralized repository of Customer Identity Data, Account Information, Advisory and Other Reference Data which otherwise was lying scattered in various lines of business (LOB). Client's various downstream applications used this data for analysis, reporting & search applications. This included understanding business requirements and analyzing data in various online applications.  Based on analysis, data models were developed and solutions for loading large volumes of data using complex data loading schedules was provided & implemented. Developed re-usable codes to reduce development cycle. Code quality was reviewed as per existing standards and that was provided by customer & best practices. The Extract Transform and Load (ETL) codes was developed using Oracle PL/SQL or Informatica and Unix/Linux Shell Scripts.