A TrendMiner Team Guest Post
TrendMiner: A Self-Service Industrial Analytics Tool Connecting Factories Around the World
In this special blog post, our very own Rob Azevedo explains how our self-service industrial analytics software connects factories at different sites both locally and globally resulting in global team collaboration.
Rob Azevedo is the Product Manager for TrendMiner. He leads the strategy and execution of building and growing the contextual analytics and dashboarding platforms launched in 2019. Based out of TrendMiner HQ in Belgium, Rob has over 9 years of experience in building software which includes B2B business platforms and high performant B2C ticketing and entertainment experiences viewed by millions. He is passionate about new user experiences driven by technology, and uses the knowledge and drive he has to bring Digital Transformation, Industry 4.0, and IoT initiatives to market.
Many industrial manufacturing organizations have multiple sites at different locations around the world. The ability to connect factories would enable better global collaboration and thus increase team and production optimization. How does TrendMiner connect factories around the world? I’ll tell you how.
#1: Access to Global Data Through the Cloud
Thanks to the advancements in software architectures and distributed setups, the connected factory is ready to take its next step. An increased number of plants are making the move from an on-prem only strategy – which still remains a viable deployment option – to the cloud. With new distributed architectures, factories from all over the world can access their data, generate insights, and share those insights on a global scale. For example, an organization’s central analytics teams of process engineers can be given access to data from remote sites, which would allow them to generate insights, propose solutions, and set up preventive and predictive measures in place. All with the help of on-site subject matter experts.
Process engineers can use their expertise and global view of the organization to deploy those insights to their entire user base in the form of recommendations, alarms, dashboards, prepared trendviews, and contextual overviews that can be used in day-to-day operations and reporting as well. TrendMiner makes a difference by fostering collaboration through the easy sharing of all process content and by providing features teams need to communicate around the world. This global collaboration lets your teams solve problems faster and better, learn from others, and run their plants more efficiently and profitably.
#2: A Birds-Eye View with Easy Search Capabilities
TrendMiner works like Google but for the manufacturing industry. And in what way? Well, it provides a cache on top of all of an organization’s data, all the different types of data (including time series from machine sensors and contextual data like quality, maintenance, and batch records) and in seconds, makes it searchable and available for analysis. In Google, there is one entry point to launch a search to every website from that point. TrendMiner is set up the same way; it is a single application that provides the tools and search capabilities to analyze all connected data points, regardless of where they come from.
The benefit to this approach of self-service analytics is that it doesn’t matter which historians your plant is connected to or how many (be it from OSIsoft, Aspentech, Aveva, or others). This allows specialists from anywhere in the world to support local teams who are analyzing process data. This approach also includes the capacity to incorporate contextual data from 3rd party business systems that can shed more light on the operational performance, such as OEE, batch, and maintenance systems.
#3: Single Production Overview Eliminating Data Silos
Typically, organizations operate in data silos, each with its specific teams and process and business applications, such as operations, maintenance, control room, and lab teams. And each has its own dataset to work with.
These silos obstruct a clear and transparent understanding of production and can lead to a slower resolution of issues. TrendMiner eliminates these communication and data silos by providing a single production overview for the entire organization, no matter where the factory is located in the world. All personnel can be given access at an administrator level to information from different factory sites which allows for global collaboration and a much faster, better, and safer resolution of production problems.
#4: Capitalizing on Process Expert & Data Scientist Skill Sets
Due to a scarcity of the data scientists whose work with soft sensors, machine learning models, and the search for data precursors of downstream problems, organizations are at a disadvantage to solve critical production issues if data scientists who typically work from a centralized location can’t access the right data.
However, over the years, self-service analytics tools have enabled engineers to be “analytics enabled”. Moreover, these analytics enabled engineers have shown an increased interest in going beyond that level of analytics maturity, further closing the gap with the data scientist in central troubleshooting groups. As an Analytics Expert, this type of user wants to create and work with notebooks that would provide an advanced analytics experience beyond the robust built-in TrendMiner capabilities, yet be seamlessly integrated within TrendMiner.
With our embedded notebooks, Analytics Expert users can load data from a time series or context view that they have prepared using the typical built-in capabilities (select a set of interesting tags, select timeframes of interest via searches, filter out data,…). They can then do some analytics automation via coding (eg. repeat analysis over a large range of assets) or create predictive tags using custom models (e.g. neural network, random forest, K-means clustering) supported by the well-known notebook libraries Pandas, NumPy, SciPy, SciKit-Learn and others.
This new technology/capability solves the problem of a lack of data scientists at a typical non-centralized plant location. Additionally, with data science notebooks, the analytics maturity of the organization will be expanded to more efficiently capitalize on both the process experts’ and data scientists’ skill sets thus promoting better collaboration locally and globally.
#5: Multi-Language Support
Process experts operate in an international context, and so does TrendMiner. Our software is available in eight languages, so you can roll out the application globally letting users have a product experience in a language they are comfortable with. This language capability results in increased use and better plant collaboration and management around the world.
#6: Blending of IIoT Technology & Existing Manufacturing Applications
The data landscape in the typical organization shows time series data that is usually stored in proven data storage solutions like historians. To get all of the benefits, TrendMiner connects factories by offering an entire solution set on top of the proven offering. 
Moreover, in the coming decade, a lot of new data will come out of green field devices, powered by the evolution of the Industrial Internet of Things (IIoT). TrendMiner blends old technology with new technology by bringing in sensor data from new green field devices – from new IIoT sensors through sources like SoftwareAG Cumulocity, Amazon Timestream or Azure Kusto. It supports new factories that might move to a modern IIoT stack, and since more remote equipment is connected at the edge, field engineers will be supported through the analytical insights of process engineers who are not on site.
#7: Contextualize and Illuminate Data
To let a global organization understand and analyze a production process, contextual data needs to be made available. For example, a process engineer from the other side of the world can see that a temperature sensor was down because maintenance was taking place at that time. Or a bad batch can be quickly identified by looking at the attached quality sample’s measurements. Experts can then use these time frames of contextual data to jumpstart or aid their next problem-solving task.
TrendMiner illuminates all data bringing time series and context data together (Figure 1). This way the right people can see and consume the information in order to execute their job even better, no matter where they are located in the world.
Figure 1. Process overview with contextual data (production data, maintenance)
#8: Showcasing Findings for Actionable Insights
Showcasing production findings for colleagues or creating dashboards that can be used globally by teams to follow up on day-to-day operations is a critical step in making the data work for your organization (Figure 2). At a glance, information needs to be consumable, and at the click of a button, users need to be able to dive into the details of that data behind the dashboard. You can do this with TrendMiner.
Its dashboards can be used to convey critical information at a glance. For example, from the input that operations teams make, maintenance teams can see the maintenance requests on their dashboards. Additionally, they can also see the health of their assets, the production alerts which are based on process monitors of different operating zones, and the outcomes of predictive maintenance models.
Shift handovers and morning meetings: everyone has them. Through dashboards, the relevant information that was generated during a shift can be bundled and displayed, so the next team can consume the information in a clean and structured way, on a digital device of their choosing. And to add to this, dashboards can be used to report key parameters to plant or upper management, so they are kept in the loop.
Figure 2. Process dashboard
TrendMiner – So Much More than Global Collaboration & Process Optimization
As you can see, introducing an analytics solution that connects factories needs to be more than a trendviewer. You need a tool that empowers your process experts, so they themselves can analyze, monitor, and predict operational performance using sensor-generated time-series data and that also sets the stage for global collaboration between an organization’s different sites and different personas.
With its one screen production overview, multi-language support, high powered search features, embedded data science notebooks, context analytics, and action focused dashboards, you and your teams around the world can use TrendMiner to share crucial process data/information for global connection and collaboration. The end result: you and your operational experts can contribute to a more sustainable, competitive, and profitable organization.
