ContextHub - An Introduction

TrendMiner's ContextHub is a detail factory, repository and search engine for Context Items that can be neatly aligned to both assets and processes. ContextHub is a platform in its own right, and can be configured to requirements, so that context itself becomes a powerful new dataset which can be both visualized and analyzed. Events can be automatically captured, manually entered or easily imported from external business solutions, such as; Laboratory Information Management Systems, Maintenance Management Systems and OEE systems, thus maximizing the flexibility of contextual information crucial to today's self-service visual analytics.

The story of your assets and processes is key to the understanding of their inner workings. The more detail provided by the story, the better the provision for understanding. ContextHub will help you to build a highly structured story around your process and assets that can be easily shared with, and searched by others.

It is up to you to ensure the creation of context, it's ContextHub's role to help you create, import, manage, access and retrieve context.

What is a context Item? 

A context Item is a block of information linked to a data point or a period of data associated with an event identified in your datasets. Instead of writing a note in a book, add your thoughts and observations to a context item. 

Why use a context Item?

Having created your context item, you can use it to compare with other context items, share insights with colleagues and add further contextual information that will strengthen understanding of your process.

Context items can be added to the graphical representation of your data, or within contextHub.

Migrating from TrendMiner Capture

TrendMiner customer success team will help support the migration from your current TrendMiner Capture setup to ContextHub. Your asset data will be migrated; and previously created annotations will be converted to Context Items.

Using ContextHub

An overview of ContextHub

ContextHub supports the contextualization (annotation) of tags, assets and attributes (components) all contained within a package called a Context Item. Context items can be used to highlight a point of information within a trendline, or used to indicate time periods of interest that have a start point and end point. Context items can be presented in a number of display modes, including the;

  • table display - a folder like list of context items;
  • Gantt display - a series of items that can represent a process; and,
  • detail display - specific information relating to your context item.

A context item can be one of many context types (batch, anomaly, temperature measurement etc.), and contains information relating to:

  • Start date
  • Intermediate dates
  • End date
  • Short key
  • Type
  • State
  • Description
  • Creator
  • Component
  • User
  • Keywords

ContextHub stores your items and has a search facility that enables you to find specific items of interest.

Searching items typically results with a list of items that fall within the frame of your search. A list of items in ContextHub is referred to as a ContextHub View. Views can be saved.

You can make the view more specific by adjusting the search parameters (filters) to include a range of different properties, like Context Type or Component. For example, if you are interested in information concerning a particular asset over a given timeframe, your views platform will enable you to conduct a search based on these parameters. You can then further filter your views to identify when your asset exhibited a certain anomaly over a period of 3 years, for example. You can then save your search (ContextHub view). All views that have been saved in the application are also listed in the Work Organizer.

Context is the flesh. To complete this analogy the bones are your assets, attributes and tags:

  • Asset: An asset is a process component, for example; a pump, a valve, a column, a heat exchanger. A number of different tags can be related to a single asset.
  • Attribute: A characteristic of an asset that has a time series signal attached to it (most of the time), for example: Flow rates, component specifications and temperatures. These are descriptive characteristics.
  • Tag: Are time series datasets that reflect asset and process operation.

To construct a new context item you will need to provide the following information:

  • Component - Tag, asset or attribute.
  • Context type - Measurement, Batch, Anomaly etc.
  • Description - Further context to your item.
  • Keywords - Related to your context item.
  • Events - Timestamp that relates to a given state of your item.

To enter a context type, you must first configure your type (you will need administrator rights). To configure a type you must provide the following information:

  • Technical identifier - Used for your imports, api calls..
  • Name - A classifier that highlights the nature of your type, eg. Measurement, Batch, Anomaly etc.
  • Workflow (optional) - Is a process path composed of states - the more common being a functional state intersected by two events; start and stop.
  • Visualization - An icon and color (context item spot). This is an image that will appear on your Trendline highlighting to you the type of context that relates to your context item.

To complete the configuration of a context type, you must first configure a workflow. To configure your workflow you must provide the following information:

  • Technical identifier - API calls, import id..
  • Workflow name - What is used when constructing a context type
  • States - The steps contained within your workflow. There must be a start and end state entered.

Once your types and workflows are configured, it will be very easy to construct new context items.

As a user you can add context to your process by choosing a component (tag, asset or attribute), selecting a context item type, and then adding further context in the form of a description and keywords.

For example:

  • Component - Asset: Reactor R302
  • Context Type - Anomaly
  • Description - Reactor overheated shortly after commencement of process, followed protocol to resolve issue. No problems since.
  • Keywords - TodoSearchRootCause

You are free to annotate your tags when using other search tools, for example, the similarity and value based search tools. Simply select your search option of choice, open the tags of interest and then click on the 3dot icon menu. You will see an annotation option in this menu, enabling you to provide context to your data signals.

You can also add an instant context item in the focus chart of TrendHub. This is a direct annotation that will result with a trendline point highlighted with an icon. This is done by double clicking in the focus chart to reveal a scooter. In the scooter dropdown you can select the speech bubble icon, which when opted for will open up a side panel for the construction of a new context item.

The following is an overview of the key elements in ContextHub:

Element

Description

Context Views

Contain context items that are logged within a time period defined in your ContextHub view’s filter. You can make the view more specific by adjusting those filters to include a range of different properties, like Type or Component.

Context Types

Context items can be grouped by different types. The types are what gives context a classification of what that data represents, and it defines the workflow that is attached to a type which the events will need to follow. The states that come with your context item workflow provide context relating to the status of an event (Process started, In Progress, Paused, Completed).

Workflow

These are a series of states you can choose when you are entering events into context items.

When creating new context types, you can configure a workflow. A workflow is a process path composed of states. Once configured, you can then create a context item and select your newly created context type. Having selected the context type, you can choose from configured states of which your event will be composed.

Events

The timestamps that intersect states; start / stop.

Note: dates in TrendMiner are represented in UTC with the option to include timezone information in alignment with ISO 8601.

States

Process conditions used as the building blocks of a workflow. Otherwise described as the periods between event timestamps; the actions, conditions and occurrences.

As administrator you will configure the context framework with types and workflows, giving users a selection of states and types to choose from when creating context. This will become the narrative (the story) around your assets and process.

Spot

The spot, also known as a bubble, is a small circular icon found on a trendline (context item spot) that flags a context type.

 

Enjoying TrendMiner? Share your experience to help the community and receive $25.
Rate TrendMiner