Explorer: Chart View

Present insights in the form of charts and under configurable metrics

charts are useful tools to understand data better, as well as present insight. Relevance AI's Explorer dashboard supports various charts. It even allows you to build up a chart view composed of multiple all configurable according to your needs.

On this page, we will explain Chart View, as well as the most common charts supported by the platform.

Add a Chart View

On the explorer dashboard, click on the + sing, select "Chart View" from the menu, type in a name for the view, chose a color for the view (optional) and hit "Add a page"

On the new page click on "Add a chart".

This will open a page with an automatically generated chart on it. Click on "Edit chart".

Note: Using "Add a new chart" button, you can always add more charts to the page.

Column chart

Bar charts are great tools to show data distribution over different categories. In this example we will categorize our data based on the categorical field "Language" (i.e. Java, Python, etc.). And show the number of times each category is downloaded by users.

The resulting chart in our experiment is shown in the image below.

Relevance AI - An example bar chartRelevance AI - An example bar chart

Relevance AI - An example bar chart

Time-series chart

To add a Time-series chart, select "Timeseries chart" under "Chart type" menu which changes the opened window to what is shown in the image below. Add the main filed to analyse in addition the desired analysis from the left side. On the right, under "Analyse over time" add the date field in the dataset and the desired interval to show the data. In our sample, we are creating a time-series chart over sentiment analysis with monthly intervals.

To present the results into categories, select a categorical field under "Categorized by". We use the field "Type" in our dataset with two values (User and Organization). when done by the setting click on "Continue". A time-series will appear on the Chart View page.

Sentiment overview chart

The is available of the Cluster/Category View. Click on "Chats" and select "Sentiment Analysis" as type of analysis on the opened window. Select the sentiment field (the field added to your dataset after running the Sentiment Analysis workflow. Finally hit continue.

Note: To generate a time series over sentiment, change the analyse mode parameter to "Time-series).


Did this page help you?