Skip to main content
All CollectionsUser Guides
Plugins: Data Profiling
Plugins: Data Profiling

Carry out exploratory data analyses to categorize variables and discover patterns, correlations, maximums, minimums, and more.

Sergio M avatar
Written by Sergio M
Updated over a week ago

Ubidots' Data Profiling transforms IoT and industrial processes by deeply analyzing time-series data, offering significant correlations and detection of outliers, to include comprehensive descriptive statistics, trend analysis, and report generation.

This is crucial for sectors where data drives operational decisions—accurate profiling can predict equipment failures, optimize maintenance schedules, and streamline production processes, thus significantly reducing costs. By providing detailed insights with no coding at all, this tool ensures companies not only save on operational expenses, but also enhance productivity through data-driven strategies.

Data Profiling is a Widget Plugin, a tool that allows you to develop widgets tailored to your needs. Once you install the Data Profiling plugin, it will become available in your widget's drawer.

Requirements

1. Installing the plugin

Go to Devices Plugins. Click on the "+" button at the upper right corner and locate the Data Profiling Plugin:

Upon clicking on the plugin, its "readme" will be displayed. After reading it, click the button to continue onto the next step. Once there, fill the corresponding form and then click on the button to install the plugin:

Now, a new plugin should appear on you Plugins module:

2. Setting the widget up

Go to Data Dashboards, and go the dashboard in which you want to use the widget. Once there, click on the "+" button located at the upper right corner and click on the Data Profiling option:

2.1. Variables for the time series analysis

Use the widget's variable selector to choose which variables you want to profile. Then, click on the "save" button:

Important note: Keep in mind that choosing too many variables can negatively impact the widget's performance.

2.2. Time window for the time series analysis

This is set through the dashboard's date picker:

2.3. Device to extract the variables from for the time series analysis

  • Dynamic dashboard:

It's selected through the dashboard's device picker:

Note: The widget doesn't work in multi-device dashboards.

  • Static dashboard:

The variables selected in the variable's selector are already the ones associated to that device in particular.

2. Generating a report

After setting up the widget, it will display the selected variables and time window:​


Click on the "generate report" button to start the profiling process. This should take up to 10 seconds for data sets containing up to 10,000 dots. Once completed, the report will be loaded inside the widget, allowing for immediate interaction and analysis.

The widget's output is a report featuring:

  • Descriptive statistics: Mean, median, mode, variance, etc.

  • Data quality: Missing values, duplicates, uniqueness.

  • Correlation analysis: Pearson correlation coefficient.

  • Distribution analysis: Histograms.

  • Variable interactions: Analysis of variable relationships.

2.1. Overview


2.2. Variable's extended information

2.3. Interactions

2.4. Correlation

Did this answer your question?