loader animation

Healthcare dataset analytics platform

AI-powered dataset management solution reduced operational costs by 29%.

Scroll down to read more

Project highlights

Industry: Healthcare
Client services: AI Product Development
Started in 2023 
Location: New York, USA 
Team size: 8 members 
Duration: 9+ months 

About the client

Healthcare consulting company specializing in data analytics and support of the medical research community, set a goal of automating its dataset management process with AI technology.

Business challenges 

Our client had developed a database designed to help health services researchers and other users find datasets to support their analytical requirements. The database included a curated list of top publicly available datasets, detailed characteristics of each dataset, and a dictionary of the variables contained within those resources.

However, it was clear that the existing system had limitations. The platform only provided metadata, not the actual data, and its search functionality was not optimized for ease of use. Additionally, the client needed to scale the database, enhance its searchability, and improve the overall user experience to better serve its target users, including researchers, analysts, and information specialists.
The client sought to transform this database into an online searchable product that could address these challenges. They required assistance in determining the ideal database type and configuration, creating queries and utilities to optimize search capabilities, and developing a user interface that would facilitate searching from various perspectives. Scalability was a key concern, as was the need to harvest metadata from multiple disparate sources.

Furthermore, they were interested in exploring how machine learning could enhance the project, particularly in improving the accuracy and relevance of search results.

Goals set to Achievion

Recognizing Achievion’s expertise in AI-driven solutions, the client engaged us to help them achieve these objectives. Our goal was to transform the existing database into a sophisticated, web-based platform. This involved automating dataset lookup and metadata harvesting, enabling natural language queries, and developing an intuitive user interface.

Additionally, we were tasked to explore how AI and ML could be integrated to further enhance the platform’s capabilities. The client’s vision was to create a product that not only served their internal needs but could also be marketed to other data analytics companies, offering an advanced solution for healthcare data searchability and usability.

Solution 

A platform that caters to medical researchers and professionals, providing access to metadata from various studies.

The service features a sophisticated data analysis tools, facilitating easy retrieval and comparison of essential data, thereby streamlining research activities.

We established the data collection mechanism to enable AI/ML models training and AI-powered capabilities implementation:

  • Intelligent data search
  • Automated metadata harvesting
  • Metadata variable summarization
  • Natural language search queries
  • Predictive analytics to determine potential gaps in the database and inform future dataset selection needs
  • Hyper-personalization of the search results according to individual user preferences

Key features of the product in detail

The platform stands as a comprehensive solution, encompassing a user-friendly website, an administrative panel, and additional functionalities.

Each component is designed to offer a seamless experience for potential users in the healthcare data domain.

User-Friendly Website Interface

The website offers an intuitive interface equipped with advanced filtering systems, keyword search capabilities, and tools for dataset comparison.

This allows researchers, analysts, and librarians to efficiently explore datasets relevant to health services research (HSR) and analysis.

Convenient Administrative Panel

Complementing the frontend, the administrative panel serves as a robust backend, employing cutting-edge technologies to manage the extensive database effectively.

This backend empowers administrators to navigate diverse dataset structures, ensure smooth import functionalities and precisely track changes.

Advanced Search Functionality 

The standout feature is the search functionality, using the Elasticsearch stemmer to provide a refined search experience.

This goes beyond basic matching, employing algorithms to identify word roots or parts, ensuring sophisticated and context-aware search results that enhance the user experience.

Synonym Support 

Furthermore, there is potential for future enhancements, including the integration of synonym support, which would further elevate the capabilities of the search feature.

The system was meticulously designed to function as an efficient and accurate data aggregator, prioritizing the user’s need to find relevant information swiftly and effectively.

Custom Analytics Integration 

The platform integrates Google Analytics for comprehensive tracking, covering both standard and custom metrics.

Notably, we’ve tailored custom analytics to align with each client’s specific requests, target metrics, and the unique characteristics of their application. This personalized analytical approach empowers clients to gain insights into user behavior, preferences, and interactions, contributing to ongoing improvements in the user experience.

Enhanced Verbal Design 

Significant efforts were also dedicated to improving the verbal design of the system. This includes user prompts, icons, and other informational cues that enhance usability and ensure users can navigate the system with ease.

The project team worked closely with the client to refine these elements, ultimately achieving an interface that is both user-friendly and functionally robust.

Optimized System Performance 

Additionally, the system’s performance was enhanced by creating paginated grids, which significantly reduced load times for large datasets. This improvement ensures a smoother and more efficient user experience when handling extensive data.

Continuous Improvement with User Feedback

The platform also includes an internal feedback system that enables ongoing enhancements based on user suggestions, continually improving its features and overall user experience.

Business outcome

Fully operational Minimum Viable Product (MVP) was deployed to production in 2024. The platform effectively combines a user-friendly interface, robust backend technologies, and advanced analytics, enabling healthcare professionals to efficiently access and analyze datasets.

The platform opened up a new revenue stream for the client’s consulting business and set them apart from the competition by offering unique approach to medical data research.

Timeline 

4 Weeks
Discovery Phase
  • Developed project architecture and infrastructure
  • Created the app’s UI/UX design
  • Documented requirements
8 Weeks 
Phase 1
  • Development — MVP
2 Weeks 
Phase 2
  • Stabilization
2 Weeks 
Final Phase
  • MVP Enhancement and Refinement

Team

Product Manager 
Business Analyst
AI Solutions Architect 
UI/UX Designer 
Backend Developer 
Frontend Developer 
QA Engineer
DevOps Engineer

Tech Stack

Backend:

PostgreSQL Server 15
Elasticsearch 8
ZomboDB PostgreSQL plugin
Keycloak
AWS API Gateway service
Java 17
Spring boot 3.1.5
Hibernate 6.2.13
Docker
Maven
Liquibase 

Frontend

React 18
Styled-components
TypeScript 4.9
React Table
React Query
Axios
Date-fns
Eslint 

DevOps

AWS
VPC
EC2
Load Balancer
API Gateway
S3
Route53
ECS
CloudWatch
Bitbucket
Bitbucket Pipelines

You may also like

Get in touch to learn how our AI powered solutions
can solve your business problem.

    *

    *

    0 from 500