December 13

  • 3 minute(s) read

Time series with automatic downsampling

When analyzing time series, if there are many points (e.g. > 100k), the graphical visualization may be not always optimal.
As the pixels on the screen are limited, a common approach is to overlap or aggregate (e.g. AVG by HOURS) data points, which may lead to a loss of information (e.g. spikes are flattened).
In addition, many points may cause browser slowdowns or even to the page being blocked. To help users to deal with large data sets, in the Time Series Chart, we introduced the LTTB (Largest Triangle Three Buckets) data visualization technique designed to efficiently represent large datasets over time by reducing the number of data points while preserving the overall trends and shapes of the data. The LTTB reduction algorithm is automatically applied to all metrics loaded without a specific aggregation, and in all cases that the number of retrieved data points is greater than a fixed threshold.
When the graph is zoomed in or the scroller is moved, the selected area is automatically expanded to display more data points, preserving memory and browser performance.

Time Series with offline highlighting  

Connected products may go offline, and in some cases upon reconnection they can publish data that were collected during the offline period.
Unfortunately, this data could be incomplete, and for a user who is looking at a chart it must be clear about where there is uncertainty in the data, this avoids making wrong decisions.
The Time Series Chart has been extended with dashed areas that highlight when the product was offline and therefore data is missing or unreliable.  

Time Series with work sessions

Work sessions provide useful information to monitor productivity, and now DPS users, like for instance Operation Managers, can correlate work sessions with measurement trends.

The Time Series Chart now allows displaying along the metric data series also the work sessions started within the same period.

Quick jump between work sessions

When monitoring the productivity, the DPS users may need to navigate among work sessions to inspect details of each one.



In the DPS whiting the work session details page, through dedicated control buttons, the user can jump to the next and previous work session without changing the page.  

Thing selection field

To improve data comprehension when consulting reporting pages that include aggregated or statistical data, it is crucial for users to focus only on the items that are relevant to their analysis. By enabling users to select multiple objects for statistical analysis, the page input becomes a cornerstone of effective and meaningful data exploration. It empowers users to save time, minimize errors, gain deeper insights, and make well-informed decisions tailored to their specific objectives.

To help managers in analyzing the products fleet, we have improved the search-field, and now it is possible to leverage the new selection-input that allows to select a specific set of things that can be used as input by other page widgets.  

Lists with custom columns

Users often deal with tables that contain a large amount of data (e.g. thing list with several property columns), much of which may not be relevant or missing for their immediate task.
Now list-based widgets provide users with flexibility in organizing and personalizing the table view to suit their preferences or task requirements.

In the template, the Console user can declare all possible columns, and for each of them define whether the column is always visible, initially hidden but selectable by the user in the DPS, or always hidden but present in the data export.

Bug fixes (Console)

  • The thing defaultName property was not available in the template visual editor.

  • The Name and Description properties of Product Model Part were not available in the template visual editor.

Bug fixes (DPS)

  • The IoT Connector configuration was not inherited by sub-things.

  • The API that allows the creation of a single thing, returned an error if the location ID was not provided to create an UNASSIGNED thing.

  • The condition on serialNumber was discarded when retrieving UNASSIGNED things.