Generate Automated Data Insight
This tutorial walks through the Automated Data Exploration feature in RATH with the Mega-auto Exploration feature.
In the following example, we will be working on the demo dataset named "Cars", which collects data of technical specs from different car brands. You can download this public dataset from Kaggle (opens in a new tab).
Prerequisites
-
Import data: On the Data Source tab, click on the Import Data button, choose Demo, and select the "Cars" dataset. To process other data sources, refer to the Data Profiling chapter.
-
(Optional) Subsetting data: On the Data Source Tab. RATH gives you a general overview of the dataset. You might omit a certain subset of your dataset from analysis by unchecking the use field option.
Start automated data exploration
-
Click on the Start Analysis button to launch Automated Data Exploration with RATH.
-
RATH will automatically process your request and redirects you to the Mega-auto Exploration tab with visual insights.
Explore visual insights
After RATH has finished processing, you may:
-
Pick a chart that from a row automatically generates visual insights.
-
Search a particular chart by inputting your keywords in the Search Views bar.
-
Add or remove a variable on the lower side of the screen.
-
Save a chart to your collections by clicking on the Star button:
Your starred charts are accessible on the Collection tab on the left side of the screen.
-
Export the chart to a PNG or SVG format image file by clicking on the export chart button.
-
Explore your data with a Tableau-like interface where you can drag and drop variables to construct charts manually. Click on the edit chart -> Customized Analysis option to edit the chart with Graphic Walker (Manual Exploration).
-
Explore your data visually with a painter-like interface where you can directly select, remove and study your data with drawing tools. Click on the data painter option to edit the chart with Data Painter.
-
Explore the automated data exploration results by checking out the associated charts. Click on the associate views button:
On the right side of the screen, RATH automatically generates relevant charts. These associated charts are categorized by Associated Measures or Associated Dimensions.
If you find a chart that fits your interest, click on the Analysis button to study the chart.
Edit charts with Vega Editor
Vega/Vega-Lite is a high-level grammar for interactive graphics, in which you can edit data visualization with declarative JSON syntax.
For more details about how to use Vega Editor, refer to the Edit Charts with Vega Editor chapter.
Best practices
-
For new databases or unexplored datasets, it is the best practice to conduct Mega-auto Exploration to get a general idea about your materials on hand.
-
If you already have some concrete ideas about your datasets, it is the best practice to move on to Semi-auto Exploration, which functions as the Copilot in assisting your data exploration journey.