Explore, trend, and analyze industrial data with Cognite Charts and instant access to data and no-code tools. You can find and work with any time series or contextualized P&ID in Cognite Data Fusion (CDF).
Cognite Charts helps you explore contextualized data from CDF to find the time series you want to plot.
Start a new chart or open a private or public chart on the Overview page. A private chart is only visible to the one who created it, while public charts are visible to the whole organization. Select the blue + button to get started.
Use the Equipment tag tab to search for the asset tag related to the time series you're interested in or the Time series ID tab to search for a specific time series.
You can also search for an
externalId (which may be different from the name) for a time series or asset/equipment tag in the search bar. Charts will automatically display the result at the top of the list. Note that
externalId requires an exact match.
If an asset is found in one or more P&IDs, you can select the icon next to the result to open the P&ID.
You can also add time series data directly from the P&ID to your chart.
Below the chart, you can see all the time series that you have added to the chart. You can rename, remove, and change units for each time series. Also, you can adjust the thickness and color of the line and select dotted or dashed lines. To show or hide a time series, click the Show / hide eye icon in the second column of the row.
Use the search icon to open the related P&IDs. Select the information icon to zoom into functions and constants.
All vertical y-axes are interactive and can be independently moved. Use click+hold and drag or scale in and out by hovering over each horizontal axis and scrolling using your mouse. This functionality enables fast stacking, overlay, and comparison of trends of multiple time series. Navigate between current and historical data with the Last year, Last 6 or 1 month, Last week, and Last 1 or 2 days buttons, or set a specific time and date in the calendar.
Note that Charts uses the local time on your computer as the default setting.
Calculations and analysis
Troubleshooting and working on root cause analysis often require calculations ranging from basic algebra to more advanced algorithms. You can apply these calculations to one or multiple time series, and the results are directly available in the chart.
To start a new calculation, click the blue + button and select Add calculation, and a new line is added with an fx icon representing the new calculation. When the node editor opens, right-click in the grey work area and add one or more inputs (sources) to get started.
By default, Automatic data alignment is enabled for all calculations, which uses an algorithm to ensure that data in calculations are automatically aligned (resampled and reindexed). If you want to define and set up your own resampling and reindexing parameters, turn off Automatic data alignment. You can manually add resampling or reindexing function nodes into any given calculation and define the parameters yourself.
Constants and functions
Right-click in the work area to add constants or functions to the calculation. Select i to view the function details or hover over the parameters to view more information.
Pipe the calculations together in the right order, including the final Output node, to complete the computational graph. This will trigger the calculation to run, and the results will be visualized on the chart plot. When a valid calculation is connected with an output calculation, the calculation runs automatically.
Whenever an input source, a parameter, or the time frame in the chart is changed, the calculation is refreshed and shown in the graph. Adjust input parameters and constants values at any time by double-clicking a function or constant.
Calculations in the list are valid inputs to a new calculation, which means you can divide the calculation steps into several separate calculations and see the output of each calculation step.
Monitor time series
The features described in this section, Monitoring and Alerts, are currently in Beta and are only available to customers via the Early Adopter program. For more information and to sign up, visit the Early Adopter Group on the Cognite Hub.
Create monitoring jobs to receive email alerts about any condition you define in a time series.
- Select Monitoring to open the panel.
- Select the time series you want to monitor in the Source dropdown list.
- Choose to monitor either the upper or lower threshold, and set the value.
- Use Minimum duration to define how often you want to be alerted. For example, only receive an alert if a breach lasts x minutes.
- Use Schedule to set how often you want the monitoring job to run.
- Select the folder where the alerts are stored, or create a new folder.
- To create a new folder, enter the folder name and select New folder.
Select Next to choose your sign in credentials, either CDF sign-in credentials or CDF Client ID and Client secret. If you don't know what values to enter in these fields, contact your internal help desk or the CDF admin for help.
Select Start monitoring.
Manage monitoring jobs
The Monitoring panel shows you the folders and the number of running monitoring jobs.
- Filter to see All folders, folders relevant for the current time series in the chart, or only folders with monitoring jobs that you subscribe to.
- Open the folder to preview the monitoring job's name, the time series's name, and the rule it uses.
- Use More options to delete the monitoring job.
- Use History to see all the alerts that have been triggered for your monitoring job.
An alert is a representation of the event triggered by a monitoring job. Depending on how many users subscribe to the monitoring job, one alert can produce zero or many notifications (emails).
All the alerts triggered by the monitoring jobs are organized in folders in the Alerts panel. A red dot on the alerts icon indicates that a new alert has been received for a monitoring job you subscribe to. You can filter to see All alerts, the alerts In current chart, or the alerts that you Subscribe to.
- Open the folder to preview the name and rule of the monitoring job that has created the alerts: time series name, time, date, and duration of the breach to troubleshoot the alert.
Alerts can be active or resolved:
- Active: no action has been taken on this alert - you must resolve it manually. Note the time series name, date, and time, so that you can find the incident on the time series.
- Resolved. When you have investigated an alert, select Active and change the status to Mark as resolved. You can also select Mark all alerts as resolved. That resolves all alerts created for that monitoring job.
Create access to monitoring and alerts
Read this file about how to create access to monitoring and alerts.
Data profiling helps you assess the quality of a time series and build trust in your data. Use the data profiling side panel to get an overview of your time series to ensure they fit your purpose. Assess how data is distributed by reviewing statistics on time deltas and gaps. Show gaps to study them closer. Depending on your use case, deviations in the pattern may indicate a lack of data quality.
- Select the time series you want to assess in the Source dropdown list.
- In the panel, you can show gaps in the plot. You can also toggle between viewing a boxplot and a histogram of time deltas or density.
Data profiling is built on the values defined in the Industrial Data Science Library (InDSL) and defines gaps if the time deltas deviate by more than 1.5 times the interquartile range.
Thresholds on historical data
Use thresholds to check whether a time series has crossed a given value while looking at historical data and for how long it crossed this threshold. You can set thresholds for checking whether a time series or calculation is Over, Under, and Between any given values. The result of the threshold, when a time series or calculation crosses a given threshold, is visualized on the chart plot.
Select Thresholds in the tools list on the right-hand sidebar to get started.
Find and visualize events on the chart plot
CDF Events are a critical and relevant resource when working with time series data in CDF. You can display events in charts to perform industrial time series analytics, whether it's a work order on related equipment or a faulty shutdown.
Select Events in the tools list on the right-hand sidebar to create a filter group to show the events that are most interesting to you. All events that match your filter parameters will be displayed on the plot for the current time range. If you move the plot and change the time range, your chart will refresh and display the updated results.
Select View results to see a list of all events that match your filter query. You can click on a particular event to highlight it in the plot. Select View details to view the full details and metadata for any event in your results.
Collaboration and sharing
CDF Charts can be private or public. You can share a link to public charts with co-workers with access to the Charts project . All users in your organization can see the charts under Public charts on the Overview page.
Only the owner of a chart can save changes made to Public charts. To save changes to a chart you don't own, duplicate the chart with the Duplicate option on the Actions menu and save changes to the duplicated chart.
You can use duplicated charts to create template charts with pre-selected time series and calculations that you can reuse when performing troubleshooting or root cause analysis and share with your team.