Client
Wúru is a health-analytics provider, whose platform leverages data and algorithms to enhance the productivity of healthcare providers and create better experiences for patients. Founded in 2019 and led by Luciano Tourn, their platform feeds on millions of data points originating from patients’ itineraries to provide healthcare workers with critical information that drives smarter decisions.
Requirements & Context
Wúru’s platform is a critical tool for decision-making in hospitals and health centers, which makes it paramount that it handles data with precision and speed.
Since the platform handles a large volume of data, coming from various sources, and the data models at the origin and destination do not always coincide, building a structured pipeline helps improve the way in which the data is processed, and the speed and quality of Wúru’s response to incidents in the handling of data.
Approach
We started by defining the technology stack that could support Wúru’s data pipeline. There is an initial instance in which the platform captures large sets of data from the hospitals, which then have to be mapped against Wúru’s data model. Depending on the results of that mapping, different options for the treatment of data open up, and the purpose of the pipeline is to optimize the organization and tracking of that process.
To implement this, we first had to redesign the structure of their data model and implement the new data model in Redshift. We used a columnar database to enable faster queries for data analytics, and opted for Airflow to visualize each step of the pipeline, log errors whenever they arose, and re-execute steps swiftly.
Results
This redesign of the data model streamlined the processing of large volumes of data, improving the overall reporting performance thanks to the use of the columnar database. Also, analyzing issues in the pipeline got much simpler, due to the observability provided by the Airflow implementation.
