EX Dashboard: Text Analysis

Unlock deeper insights from your employee feedback with the enhanced capabilities of Text Analysis, powered by OpenAI. While metric questions are essential for understanding employee experiences, exploring the reasons behind the ratings is equally important. Whether feedback is positive or negative, open-ended responses provide valuable insights that help shape decisions and drive meaningful improvements.

Text Analysis enables you to identify common topics and analyze positive, neutral, or negative feedback with ease. Drill down into key segments to understand what drives employee sentiment and take informed action.

Navigate to

• Open-ended Responses
• Theme Analysis
• Text Classification

Open-Ended Responses

This tab includes the following widgets:

• Overall Sentiment
• Sentiment Trend
• Word Cloud
• Response Sentiment
• Responses

Overall Sentiment

The Overall Sentiment widget, displayed in the top-left corner, shows sentiment distribution using a Donut chart categorized into Positive, Neutral, and Negative. Hovering over a sentiment reveals the total number of responses, percentage share, and the percentage change compared to the previous day, month, or year.

The three-dot menu in the top-right corner of the widget has the following options:

• Graph type: This option allows you to change the graph type of this widget. The available graph types are Donut Chart, Pie Chart, Bar Chart, Column Chart, Area Chart, Horizontal Stacked Bar and Vertical Stacked Bar.
• Share: This option allows you to share access to the real time data of this widget with anyone who has this link.
• Download: This option allows you to download this widget as a PDF or an image file(png and jpeg).
• Refresh: This option allows you to refresh the data of this widget.

Sentiment Trend

This widget displays sentiment trends using a vertical stacked bar chart across daily, weekly, monthly, quarterly, or yearly intervals. Hovering over any data point shows response counts and percentage distribution for each sentiment, along with percentage change compared to the previous time period.

The three-dot menu in the top-right corner of the widget has the following options:

• Calculation: There are two types of calculation methods available:
  1. Cumulative: Cumulative refers to a calculation method where data is aggregated over time, meaning each data point includes all previous data up to that point.
  For example:
  January: 100 responses, 80 positive → 80%
  February: 50 responses, 30 positive

  Therefore, Cumulative for positive responses till February is,

  Total responses = 150
  Total positive = 110
  Positive % = 73.3%

  So, February reflects January + February combined, not just February alone.

  2. Non-cumulative: Each point on the graph represents values collected within that specific time range only (e.g., that day, week, or month).
  For example:
  January: 20 positive responses
  February: 30 positive responses
  The chart will show 20 and 30 separately, with no carryover from earlier months.
• Graph type: This option allows you to change the graph type of this widget. The available graph types are Donut Chart, Pie Chart, Bar Chart, Column Chart, Area Chart, Horizontal Stacked Bar and Vertical Stacked Bar.
• Frequency: This option allows you to configure the X-axis as Daily, Weekly, Monthly, Quarterly, or Yearly.
• Scale Range: Choose between Original Range or Best Fit.
• Share: This option allows you to share access to the real time data of this widget with anyone who has this link.
• Download: This option allows you to download this widget as a PDF or an image file(png and jpeg).
• Refresh: This option allows you to refresh the data of this widget.

Word Cloud

The Word Cloud highlights terms based on their frequency in free-text responses. Larger words indicate higher usage. The three-dot menu lets you change the word type, adjust word count, refresh data, or share the visualization. You can choose between displaying Emotions or Themes. Clicking a word lets you drill down into related responses and associated sentiment.

Response Sentiment

This widget shows response distribution over time, categorized as Positive, Neutral, or Negative. Hovering over a sentiment displays intensity, themes, topics, and the response text. The three-dot menu allows sharing, downloading, or refreshing the data.

Responses

All open-ended responses are displayed in a table showing sentiment, themes, and response date. You can update the sentiment of a response using the dropdown in the Sentiment column. The three-dot menu allows column customization, sharing, downloading, or refreshing data.

Theme Analysis

Theme Analysis identifies recurring topics in employee feedback by analyzing language patterns and detecting common keywords and phrases. This view provides quick insights into employee priorities, concerns, and successes. Drilling down into themes offers deeper visibility into specific discussion points.

Available Widgets

• Negative Themes: Displays negative themes using a Pack Chart.
• Positive Themes: Displays positive themes using a Pack Chart.
• Theme Word Cloud: Displays identified themes
• Emotion Word Cloud: Displays detected emotions

Text Classification

The Text Classification tab is the default view that provides key insights into your open-ended responses. The data displayed here is generated independently and does not rely on AI services.

Available Widgets

• Overall Sentiment Score: This view shows a selected topic’s overall sentiment with a breakdown into Positive, Neutral, and Negative responses.
• Response Sentiment: Displays positive themes using a Pack Chart.
• Responses: Displays identified themes

Overall Sentiment Score

All responses will be categorized into topics. Hover over each donut to view detailed insights, including:

• Number of responses categorized within the topic
• Mean sentiment
• Median sentiment
• Number and percentage of positive responses
• Number and percentage of negative responses
• Number and percentage of neutral responses

Response Sentiment

This widget shows response distribution over time, categorized as Positive, Neutral, or Negative. Hovering over a sentiment displays intensity, themes, topics, and the response text. The three-dot menu allows sharing, downloading, or refreshing the data.

Responses

This table displays all open-ended responses along with their associated sentiment, identified topics, and the date and time each response was received.

The sentiments and topics are automatically analyzed and assigned by the system. If you believe any sentiment or topic has been categorized incorrectly, you can update it manually. Sentiment can be modified using the dropdown menu next to the sentiment score in the response table. Additional topics can also be added from the system-generated predefined list by clicking Add Topic. The three-dot menu provides options for column customization, sharing, downloading, and refreshing the data.