Using the latest Big Data innovations, bigCONNECT stores unlimited information, understands it and makes it available for your discovery.

Information can be from any imaginable kind of source - documents, databases, images, sounds or movies. bigCONNECT contains intelligent algorithms that process your data to infer new knowledge. It processes and understands your data, then it shares it all with you.

Using data
to your needs

The platform can continuously draw in more data from publicly available sources, correlate it with your existing data and recognize patterns. Using a real-life model of interpretation, the large volume of data is automatically correlated and presented to users using graphic representations, very close to human thinking.

The Benefiters

If you are a FMGC Company – It can be used to better understand the market for your product. Are you a government defense organization – it can help efficiently deploy troops in a combat area. It can supervise the client base for a Telecom Company, or correlate specific information from thousands of reports. Find out what bigCONNECT can do for your business.

How it works

bigCONNECT first acquires the data from all possible sources. It then enriches
it, by adding markers and tags to it, and finally displays all the new


Collect data

bigCONNECT can ingest and process any kind of information. Whether you have it stored in databases, office documents, text files, XML files, HTML files, images, audio files, video files or different streams, Facebook, Twitter, YouTube, Google+, and web sites, our content extraction engines and crawlers will automatically extract the relevant information as objects, relations and attributes.

Database ingestion

bigCONNECT can connect to any kind of database, extract information through queries and load it in the knowledge graph using our intuitive mapping editor.

Structured files ingestion

Just drag an XLS or CSV file on the workspace and bigCONNECT will automatically parse the content and allow you to configure the mappings in our mapping editor.

Unstructured content

bigCONNECT can process most types of documents to extract and correlate relevant information such as objects, relationships and attributes. Metadata is extracted from images and the OCR engine will extract all text information which is further analyzed and correlated with existing data. Audio and video files are processed by the multimedia pipeline: audio files are processed by the Automatic Speech Recognition engine and video files will be split into images and audio tracks which will be again processed by the content extraction engines.

Web Crawling and Social Networks

Data from external sources such as web sites and social networks can be scraped and mapped to the semantic model using Sponge.

Enrichment - Bring data to life

Trained Natural Language Processing (NLP) engine tokenizes your text down to semantic level and place resulted information in the right shelf. By following a sequence of steps inspired from human thinking, the NLP engine first understands the context then identifies key concepts, their properties and their networks. Manual content enrichment further traines the engine improving its intelligence. Detection can be precise, using preconfigured dictionaries, or can be self-trained, using Maximum Entropy models.

bigCONNECT Voice

Enables you to train your Speech Recognition engine for any domain, any language, any dialect. The training requires dictionary definition, phonetic configuration and audio samples and is executed using Deep Neural Network models.

bigCONNECT includes players for audio and video content, with support for transcriptions of both audio and video files. Images can be attached to any object in the system and can be tagged to mark areas that contain new or existing objects.

The enrichment engine contains pluggable and configurable endpoints for integration with any kind of Speech to Text engine, OCR engine, Face Detection engine, Face recognition engine, image tagging and analysis engine, on premise or in cloud.

Discovery - Understand rather than read

Any information stored in the semantic model can be easily found using our semantic Enterprise Search engine.

Search interface

The search interface is designed to offer the capability of discovering any information with just a few clicks. The refinements panel can be used to quickly drill down inside the different attributes of interest from the semantic model. Text suggestions are available to suggest relevant content words while you type. Advanced searches can be performed using our intuitive Query Builder.

You can create complex search queries using filters and groups of filters that can be combined using Boolean expressions.

Full text search

Full text search capabilities allow you to dig deep in your content. Fuzzy search enables you to find likely relevant items even though search argument words and spellings may not exactly match. Proximity search retrieves items containing two or more separately matching term occurrences located in the text within a specified distance. All this in milliseconds for billions of records.

Advanced graph queries

Enable you to discover patterns and behaviors in graphs visually, using a close to human language syntax. They overdraw the limits of traditional structured databases discovery allowing understanding data rather than reading data.

The search engine is complemented by our spatial geo-searching engine to allow finding information in specific regions on the map simply by drawing a circle on the area of interest.

Graph visualization

The graph visualization allows you to discover relations between entities, related items, similar items and to discover if and how entities are related. Page rank, betweenness, eigen centrality, degree centrality and much more are included graph algorithms that provide valuable insights from data correlation. Find influencers in a social network, group similar items, find out how objects are connected, find the shortest path from an object to another.

The temporal distribution on data can be analyzed in the timeline aggregator which can quickly position and restrict your investigation in time, at a one-click distance.

Data with geographical dimension can be analyzed on a map. See where most things happen and add geographic layers as needed to get better insights. Information can be aggregated based on multiple properties and presented as heat maps, thematic maps and regional analysis. The map includes a complex set of tools that can be used to browse and enrich information using geometry objects such as points, lines or polygons.

Analyze – better understand your data

How you see your data is how you understand it.

Asses your information. Define behaviors, angles and observe.

The way you store your data is critical, but what you learn from it is even more. bigCONNECT allows you to conceptualize your data and define a digital consciousness.

In order to fully assess your information and to understand its meaning from your point of view, bigCONNECT introduces the concept of behavior.


Behaviors are collections of queries - attribute based queries, spatial queries, graph queries, full text queries - associated with weights. An entity is subject of a behavior if its score bypasses the behavior threshold. These indicators allow you to define and balance the queries meaning. Behaviors can be scheduled to alert you when subjects change, enforcing a proactive observational dimension.

Data aggregation

Offering 360-degree visualization of data, bigCONNECT very easily defines user dashboards using powerful visual components like pie charts, line charts, word clouds, timelines, scorecards and more. With one click, the engine contextualizes your dashboard, quickly adjusting your observing angle.