Healthians is India’s leading health test at home service offering a wide range of health tests across 250+ cities of India, and counting. It has a network of state-of-the-art fully automated laboratories and a large team of highly skilled phlebotomists who specialize in sample collection from homes.

Technological innovation has been the foundation of Healthians in the industry. The company has completely turned a customer’s diagnostic experience on its head, by focusing on delivering the best service experience fulfilled through technology.

Healthians was facing challenges with data growth as it was not supported by existing data platforms. Also queries became slower due to increased data volume. Due to this Metabase BI dashboards faced performance issues and real-time business insights were challenging. The existing machine learning models were limited to static feature variables handled manually and not scalable.

Existing data ingestion was happening from multiple data sources through API and third party software. Healthians was using in-house built manual recommendation systems to support their product sales. Searce had recommended a scalable, intelligent advanced data analytics platform which includes real time data replication from sources into Data Lake and a data driven product recommendation engine to modernize their current ML model built on Sagemaker.

Data sources consisted of:

• Applications, consumer apps for Android and iOS, websites, Partner apps, CRM and DocumentDB
• Time series data is also there which is stored in Clickhouse
• Backend data is on AWS RDS Mysql DB. Audit trail CRM data and History data is part of RDS

The Searce solution involved:

• Searce used AWS DMS to load the historical data and near real time data from RDS to S3
• Searce developed AWS Glue jobs to load the data from DocumentDB, Couchbase, Clickhouse to S3 without transformation in batch mode
• Searce developed transformation jobs in AWS Glue to process RAW data from S3 and load processed data in Redshift tables for reporting
• Searce developed data-modeling as per business data domains, KPIs
• Dashboards were developed on Amazon Quicksight using wireframes and further dashboards were embedded into Healthians Portal
• With the help of data in data lake, Searce developed custom model for recommendation engine
• Searce trained the machine learning model on AWS Sagemaker and compared their performance using suitable evaluation metrics

With the new Architecture design Healthians was able to:

• Centralized the data in AWS Cloud and modernized the data warehouse in Redshift where various data sources were integrated
• Better decision making and real-time insights based on the reports in QuickSight
• Successful implementation of Healthcare Test Package Recommendation Engine with the confidence level >80%

Searce’s solution enabled Healthians to modernize the data analytics and machine learning platform to reduce the operational overhead with easy to scale on the company’s data growth. The implementation of recommendation Engine for Health Test packages improved the business performance and decision making.

Industry: Healthcare
Workload: AWS Networking Components, Amazon S3, AWS Lambda, AWS Lake Formation, AWS Glue, Amazon Redshift, Amazon Quicksight, AWS Sagemaker, AWS DMS