Methods for augmenting and accelerating data visualization, using a vector tile map, with metadata from an external database

Abstract

Vector tile maps give us new and innovative ways of using maps to visualize data. The ability of the user to dynamically change and interact with the map can lead to interesting new user experiences. In this thesis we showed ways of using vector tiles as a database, a solution we call VectorTileDB, by storing data as metadata of features in the vector tiles. This data is fetched from the tile server whenever the viewport of the map changes. We tested how well VectorTileDB performed against an industry standard NoSQL database MongoDB. We also explored what technologies are used for creating, styling and serving vector tiles. The vector tiles used were in the Mapbox Vector Tiles format, which uses Google’s protobuf technology for encoding data. This leads to a very fast and efficient fetching of data, that scales very well. There are still many questions concerning the use of VectorTileDB as a production ready database that were not covered in this thesis. But this work shows that this technology is something that has potential and should be further researched. The thesis also explored how to efficiently use the data from the VectorTileDB to augment a map, using either the client-side map library MapLibre or using an external library called Deck.gl to render data. Both solutions proved to work well with VectorTileDB.