use case
Having medical hardware and software solutions used by users worldwide, it was difficult to have a centralized overview of the most or less used features, workflows, or materials, and decisions on how to further extend the solution in order to get more business value were not correlated with the real metrics and device usage and user behavior.
During an assessment of key metrics relevant to our customer’s needs and understanding of the business processes implemented in the desktop solution apps, we assembled a team of two data analytics engineers and one business analyst. Their main goal was to transform raw data into insights.
Over six months, the team developed ‘Data Analytics’ – a business intelligence solution that ingested data from multiple sources, including logs from desktop applications, databases, and APIs. This was followed by data transformation processes in Logstash for cleaning, enriching, and ensuring data compatibility with Elasticsearch. After indexing and optimizing performance on the search engine, interactive and user-friendly dashboards were designed based on stakeholder requirements and key performance indicators (KPIs).
The different visualizations provided fast and real-time data insights about the products and user behavior patterns, serving as an instrument that drove business decisions.
PowerShell Scripts will fetch data from the Azure Blob Storage (via Analytics Downloader) and it will push data via TCP into Nginx.
Nginx instance will forward the received events (logs) and store them into Redis database, with persistence activated while Redis – used as a buffer zone (message broker), in order to facilitate smooth streaming even through variable throughput loads (events spikes).
ELK Stack – Logstash will ingest events (logs) from Redis and will run a custom pipeline, (parse, filter, geoIP) and will push the transformed data into Elasticsearch cluster with multiple instances to index the data and Kibana – for data visualizations and dashboards.
