Semi Structured Data

This kind of data is a bit abstract to imagine. If we don’t come from computer science background, let’s imagine that this type is somewhat like a table in Excel file, but the cell can have another table inside. Keep reading for further explanation.

Minimum Viable Product (MVP) Story

This data type is a combination of structured and unstructured data. The prominent usage is mostly on web and mobile app development. When the engineers built for the first version, both web and application have limited features, something that we called a Minimum Viable Product (MVP). This MVP will enable the communication using a method, for example an API (Application Programming Interface). Our devices (smartphone, tablet, laptop or PC) will send this API request and the servers will response it in JSON or XML format.

Next Development Phase

In later development, the engineers may add more features. Now the same API call we made earlier will return more data. This is where the semi structured data appear. We send the same request (structured) and we get more information (unstructured). The response follows the convention of key-value pair. But within the value, there can be another key-value.

For example, we want to provide an address. In first version, the key will be address and the value is the full address (street name, district, city, country, zip code). Later on we found this is an issue because users can put any text they want and we can’t get an accurate data. Then, we change this approach. Using the same key (address), we will have another keys inside (street name, district, etc). And these new keys will have their own value which reflect the data correctly.

Closing

This approach provides a scalable and lightweight mechanism to enhance the application we build since we can add more data, change the structures, the device can select which data is useful to be displayed, and the servers can still store the data.

See you in the next post!