Request a call
  • Hidden

Business Overview

Frame 2230422

Business Overview

The client is a large scientific publishing house and provider of remote postgraduate education with 3D visualization content for those who need advanced training courses.

The company had a flagship product—an educational system that provides web, tablet, and phone user experiences and allows users to study human anatomy in 3D.

This solution is used by leading medical universities worldwide with learning tools covering more than 13,000 anatomical structures.

Challenge

The client’s product captures the customer’s usage data using two ways:

  • App data: includes click events, visit frequency, session activity, etc.
  • Server-side data: includes license/purchase info, server-to-server notifications, and other transactional data.

The volume of gathered data was increasing and the existing solution couldn’t cover business needs. 

The client approached NIX for a solution from scratch that would provide actionable insights regarding sales and marketing performance—including in-app and email analytics—and point out ways for enhancements. The NIX data team was responsible for developing and integrating a data pipeline, creating and maintaining BI dashboards, and working on ad-hoc requests. 

Frame 2230395 (6)

Solution

The NIX team set up data storage and built a BI reporting solution on top of the AWS cloud using Snowflake as a data warehouse and Tableau as a visualization tool.

The solution has Infrastructure as Code (IaC) that decreases discrepancies, enables full traceability of each configuration’s changes, and makes the entire software development life cycle more efficient. Used services give options to process data on schedule when needed (ex. once a day) or have real-time analytics.

Based on the product backend database, third-party integrated systems, and email streaming data, we set up automatic processing of analytical data and output it on demand in a clear and easy-to-use way. 

All stages of data flow have a setup monitoring in place, which notifies the team in case of any issues. On top of that, we implemented data quality checks to track different anomalies.

  • 01

    Data collection

    We collect data from operational databases, cloud storage, and services using database migration service (DMS), AWS Lambda, and custom solutions ingested into the Kinesis stream.

  • 02

    Data processing

    After the data is ingested into the Kinesis stream, a containerized consumer processes it to remove any personally identifiable information (PII) and shape it into a common structure before saving it to the recovery bucket and ingesting it into the data lake staging area.

  • 03

    Modeling & Analytics

    The solution for data modeling and building a data warehouse is based on Snowflake. Data is moved from tables from numerous sources in the data lake staging area to the data warehouse.

  • 04

    Visualization

    Dashboards are built in Tableau, providing users with interactive analytics—the number of subscriptions sold and renewed, KPIs, in-app user activity such as duration and number of sessions, and more. Users can filter the data, drill down to different cross-sections, compare different periods, etc., receiving actionable insights.

Outcome

The developed AWS-based data analytics solution on top of Snowflake and Tableau helped the client compose a holistic marketing strategy and empowered product managers with knowledge of which areas of the application were underutilized by users and where enhancements should be made. The new data solution improved the data processing and its quality as well as expanded the list of metrics for a more thorough analysis.

As a result of the constructed solution’s success, the customer opted to expand the data platform to connect data from more applications. New scalable architecture and a robust data pipeline developed by the NIX team make it possible due to reusable components that simplify maintenance and harmonize operations, along with a unique data gathering layer depending on the various types of data the product is creating for analytics.

Frame 2230424 (2)
Team:

Project Manager, 2 Data Engineers, BI Developer, Python Developer

 

Tech Stack:

Python, AWS, Snowflake, PostgreSQL, MongoDB, Tableau, PySpark, AWS Kinesis

Contact Us