Strictly speaking, MongoDB is a document-based, database system with impeccable scaling and flexibility required for multiple levels of querying, indexing, etc. MongoDB's efficiency is renown across the board as it is a favorite of major companies like Uber, Lyft, Amazon Web Services, and more.
A little bit about the history of MongoDB: Mongo was created in 2007 when Doubleclick , an advertisement company that worked with around 400,000 advertisements per second faced tremendous issues using the database systems available at that time. Those types of systems are known as relational databases, meaning their storage methods require and provide access to data points that are related to one another. A good example of a relational database is Excel Spreadsheets, which was a popular database that used columns and tables to store data. Each storage container usually has a value for one or a few attributes which created a working relationship among data points. However, this was not the case. Overtime the storage methods of relational databases made it difficult to add information in already-bloated spreadsheets, making it difficult to access data due to massive time complexities.
MongoDB was thus birthed out of the inefficiencies of its predecessors. The programmers of DoubleClick took on the task to create a radically different foundation for MongoDB's storage systems instead of using the usual relational database systems taken on by the modern systems of that time. MongoDB utilized a document-based database. At its core, it's construction consisted of JSON-like documents which were later organized into collections.
Without getting too deep into the computer science aspects of what that accessibility looked like, the main takeaway from this overall picture being painted here is that within a document-oriented database, the capacity to store data within their own documents paved the way into the high level scalability and flexibility that we see today. Along with this, MongoDB has the capacity to stream across multiple servers at once, paving the way to the massive sociability and geospatial querying used on a daily basis.