India's health dataset is messy and contains many pieces of contradictory information. E.g. eye clinics that offer only pediatric services, hospitals located outside of India listed, among other quality issues. We cleaned up obviously wrong hospitals (~2% of facilities data), and for the remaining rows we came up with a quality score + list&map interface with two key dimensions:
- Is the data trustworthy? Yes/no/needs review
- Is the hospital Surgery-capable? Yes/no
We provided an insight tool (list view + map view) into the trustworthyness of the data - is data approved / unacceptable / needs manual review. This enables to filter out data errors and leads to quicker and trustworthy decision making for .
We performed extensive cleaning of the data and provided an interface to search for the name of a hospital with data trustworthyness booleans. Goal is to have a map for easy glance into the geographical data.
Built With
- databricks
- genie
- python
Log in or sign up for Devpost to join the conversation.