Get More Out of Cognite Data Fusion

Cognite Data Fusion is a powerful platform for managing and contextualizing industrial data. We have been using Cognite Data Fusion (CDF) for some time now, and the system is a great one. The service does have some drawbacks, but we would say only one is a major one; the rest can be remedied with proper extensions.

Ease of use is always appreciated, and it's no different here.

Oftentimes, developers of professional tools and systems deprioritize the user experience and only focus on the offered functionality. Function then wins over form, when the two are supposed to be on par with each other.

At Order Group, we put emphasis on a positive user experience, and we like to polish our interfaces. Iza Mrozowska, our Head of Design at the time, made sure everything looked up to par.

As for Cognite Data Fusion, it is quite easy to use, and technical documentation is there to help you get started with CDF.

Extensive Documentation and Training

Having extensive technical documentation helps when you're starting out or when you have an issue. CDF's documentation gives you information on:

• How to Search for Data
• How to Analyze Data
• CDF's Canvas
• CDF's Charts
• Integrations with Grafana, Excel, and Power BI

That's not all. There is also a whole section dedicated to data engineering, as well as CDF Admin. The articles go deep enough to answer all the most common questions.

There is also a link to the documentation on Cognite's Python Industrial Data Science Library.

Flexible Data Extractor Deployment

Data extractors push data in the original format to Cognite Data Fusion. It's the second step in the pipeline, after data sources and before the staging area.

There are several pre-built ones that are super easy to deploy. They are often available as a Docker container or even as a Windows .exe file. Pre-built extractors include:

• Cognite DB extractor
• Cognite PI extractor
• Cognite PI AF extractor
• Cognite SAP Extractor
• and more

Publishing docker containers makes deploying an extractor effortless. You may then simply deploy it to Azure, AWS, or Google Cloud in a matter of minutes. It's also a breeze to deploy it on your infrastructure, should the need arise.

If prebuilt extractors don't cut it, you may easily build custom ones. Cognite maintains two SDKs: one for Python, and another for .NET. For more information, go here.

Robust API

Beyond the SDKs, CDF exposes its data through a REST API, so custom extractors, dashboards and integrations can be built on top of it.

Advanced Technical Support Is Required for the Initial Setup

There are relatively few cons, though one of the biggest issues with CDF is the time it takes to get started. Unfortunately, as the case of one of our partners showed, it may even take you 3 months to do so.

It's not the worst, since you may purchase additional support, although that's not the best consolation. You're paying for the product while it's still getting set up, and on top of that, you (must) pay for the advanced support.

We use Cognite Data Fusion to collect, process, and contextualize data from Kyoto's energy storage systems. This data includes:

• Temperature
• Pressure
• State of charge
• Power output

We use this data to monitor the performance of KYOTO's energy storage systems and to identify potential problems. Thanks to the data contextualization (on top of analysis or any other operations), we can extract insights and also display them on top of specific parts.

While Cognite Data Fusion excels at data collection and contextualization, Order Group sees opportunities to do more for users, particularly those in operational and management roles.

Here's how Cognite Data Fusion can be extended. It's worth noting that the bare functionality of CDF will be enough in some cases, but for more specialized uses it definitely won't be.

Enhanced Operator Tools for Real-Time Decision Making:

Operators require clear, customizable dashboards that present critical data points in a user-friendly format. Real-time anomaly detection and actionable alerts can help operators spot and address potential issues amid constant noise, before they escalate.

Improved Search Functionality for Faster Information Retrieval:

The current search functionality can be improved to understand the context of user queries. This means recognizing not only keywords but also the relationships between different data points. Imagine searching for "battery pressure" and getting results that include related data like temperature readings and recent maintenance logs.

One way to do that is to integrate Natural Language Processing (NLP). Allowing users to search using natural language can significantly improve search efficiency. Operators should be able to ask questions like "What was the average battery temperature last week?" and receive relevant results quickly.

Reducing cognitive load will let your employees work more efficiently than before. They have a finite amount of energy per day, which they have to use as wisely as possible.

Creating an All-Encompassing Experience for BMS & EMS

Finally, together with KYOTO, our goal is to build software that handles all aspects of battery management, from initial order placement to smooth delivery, installation, and battery usage monitoring.

We want this process to fit into a larger journey, so customers have a hassle-free experience from start to finish. Our goal is to streamline the entire battery management process and remove stumbling blocks along the way, so our customers can rely on a dependable and efficient system that meets their requirements.

Cognite Data Fusion is a powerful platform for managing and contextualizing industrial data. Order Group has been using Cognite Data Fusion for quite some time now to manage the performance of Kyoto's energy storage systems.

We believe that Cognite Data Fusion has the potential to be an even more powerful tool if it is augmented to offer more features and functionality. If you would like to talk further about adding more functionality, do not hesitate to contact us now.