introducing

cloud

empire

MULTI-DIMENSIONAL

DATA LAKE

Cloud Empire is our multi-dimensional data lake - a proprietary store of New Aerospace information. We collect different types of data that are in a wide variety of structures and formats. By holding it all in a data lake and extracting meaning from it using machine learning we are able to query it in novel ways, uncovering links and insights that otherwise would not be possible to uncover.

​

We use Amazon Web Services to host the data and a range of services to extract meaning from it in different ways - we're building a knowledge graph for New Aerospace intelligence.

THE PROCESS

DATA COLLECTION

DEFINE & CRAWL

We understand aviation and aerospace and are familiar with the unique language used in the business. This is crucial in building highly-targeted keyword sets.

​

We use these keywords to set our crawlers to work, collecting and storing the most relevant data in Cloud Empire.

CLEAN & STRUCTURE

MAP SCHEMAS

All data contains noise and requires a clearly defined structure. Once we store data we then set Cloud Empire crawlers to work to define schemas for it - mapping the structure of each bucket of data.

​

This is essential and highly skilled work that enables machines to make sense of the information that is stored and query between different databases of disparate data.

ENTITY EXTRACTION

NLP CLASSIFIFERS

Different datasets don't obviously relate to one another in raw form, but by training a set of specialised algorithmic classifiers we can extract meaning and discover non-obvious links between entities.

​

As an example this might mean automatically pulling company names or aircraft types from a news article.

a knowledge graph

CUSTOM ONTOLOGIES

Once we store and classify data the process of building a domain-specific ontology for New Aerospace has begun. By defining categories  identifying and categorising objects and their properties we can start to uncover relationships between them.

​

Our long-term aim is to have a store of data that can be interrogated with natural language questions and yield meaningful insights.

DATA WE'RE COLLECTING NOW