DATA

1. Data capture

• First your specific data is captured;
• This data could, for instance, stem from sensors;
• Specialized partners capture this data individually.

2. Ingestion

• When data is generated Sirus enters the picture and sets up a streaming platform;
• By setting up this data platform your data is captured in real-time from event sources like sensors, mobile devices, cloud services or software applications;
• This streaming platform streamlines your data pipelines and makes sure the right information is at the right place for later interpretation or retrieval;
• We work with secure and scalable real-time data ingestion services like Apache Kafka or Azure Event Hubs.

3. Harmonization

• We set up a NGSI LD interface or a specific data model from the client;
• In this phase we standardize the different data formats in order to combine data from different sources;
• By standardizing the data, we provide users with the ability to compare different sources of data;
• We integrate using components like Apache Nifi or Azure Stream analytics.

4. Output

• Now we set up the Open Data Platform. Users with permission to this platform can access collections of open data from your company or city;
• The data on this platform is easily accessible. The user-friendly platform provides a clear overview of important data;
• The platform can include context brokers like Scorpio which allow data to be managed in a decentralized manner, data catalogs such as CKAN or tailor-made solutions adapted to customer requirements.

5. Analysis & Intelligence

• We work with Digital Twins which renders a virtual representation of the (current) state of your city, company or production process;
• Digital Twins is used to model even the most complex environments and make them scalable (e.g. a city or country region). It helps break down silo’s so you can gather valuable insights from your business processes in an easy and comprehendible way;
• Next, we implement Machine Learning. Subsequently, when the platform is exposed to new data, it is able to independently adapt by automatically retraining and validating the models using MLOps. In other words, your system becomes self-serving;

• By implementing Machine Learning we can also build a statistical representation of the system so we can detect data points that deviate from a dataset’s normal behavior (anomaly detection). This way we can detect faulty or drifting sensors, improve data quality or alert about abnormal or potentially dangerous situations;
• We answer your business questions and meet your requirements using the insights AI models can provide about your data. We can make predictions, build recommendations, trigger warnings and automate actions;
• We work with components like ML Studio, Python and/or Power BI for the analysis of your data.