Analytics Dashboard

Build custom analytics dashboards in PromptOwl with AI-powered visualizations, team filtering, and conversation insights.

This guide explains how to use the Analytics Dashboard in PromptOwlarrow-up-right to create custom visualizations, track metrics, and gain insights from your AI conversations.


Table of Contents


Dashboard Overview

The Analytics Dashboard provides a comprehensive platform for monitoring AI conversation data with customizable visualizations.

Key Features

Feature
Description

Custom Cards

Create personalized analytics visualizations

Multiple Chart Types

Bar, pie, line, word cloud, tables, metrics

Team Filtering

View data by team or across all teams

Date Ranges

Filter data by custom time periods

AI-Powered Analysis

Dynamic cards using LLM for insights

Drag-and-Drop

Easy page customization

What You Can Track

  • User engagement and activity

  • Conversation topics and patterns

  • Satisfaction scores and feedback

  • Token usage and costs

  • Model performance

  • Custom metrics via AI analysis


Accessing the Dashboard

  1. Click Analytics in the left sidebar

  2. Or navigate directly to /analytics-dashboard

Screenshot: Analytics Navigation

First-Time Setup

When you first access the dashboard:

  1. Default pages are pre-configured

  2. Standard metric cards are available

  3. Team filter defaults to your teams

  4. Date range defaults to last 30 days


Understanding the Interface

Dashboard Layout

Interface Elements

Element
Location
Purpose

Team Filter

Top left

Filter by team

Date Range

Top right

Set time period

Create Cards

Header

Build new cards

Edit Page

Header

Customize layout

Page Tabs

Below header

Switch between pages

Cards Grid

Main area

Display visualizations


Creating Analytics Cards

Method 1: Card Builder

Create custom analytics cards with AI-powered analysis.

Step 1: Open Card Builder

  1. Click Create Cards in the dashboard header

  2. Card builder interface opens

  3. Configure your card settings

Screenshot: Card Builder

Step 2: Configure Card Details

Field
Description
Example

Title

Card name displayed

"User Satisfaction"

Description

What the card shows

"Weekly satisfaction scores"

Data Source

Analytics prompt to use

Select from available

Step 3: Choose Visualization

Setting
Options

Visualization Type

Number, Bar, Pie, Line, Word Cloud, Table

Card Size

Small, Medium, Large, Wide, Tall, Extra Large

Table Rows

5, 10, 20, 50, 100 (for tables only)

Refresh Interval

Real-time, Hourly, Daily, Weekly

Step 4: Save Card

  1. Click Save Card

  2. Card saved to your collection

  3. Add to any dashboard page

Method 2: Page Editor (Drag & Drop)

Add existing cards to pages quickly.

  1. Click Edit Page on any tab

  2. Modal opens with available cards

  3. Drag cards to the dashboard grid

  4. Resize and reorder as needed

  5. Click Save Page

Screenshot: Page Editor

Standard Cards Available

Pre-built cards ready to use:

Category
Cards

Users

Total Users, Active Users, Top Users

Engagement

Average Messages, Time on Site

Performance

Satisfaction Score, Response Time

Usage

Total Tokens, Most Used Models

Content

Topic Distribution, Recent Activity


Visualization Types

Number/Metric Display

Best for single key metrics.

Use Case
Example

Percentages

"85% Satisfaction"

Counts

"1,234 Users"

Scores

"4.5/5 Rating"

Data Format: Single value or percentage calculation from binary data (1/0).

Bar Chart

Best for comparing categories.

Use Case
Example

Category comparison

Topics by frequency

Rankings

Top 10 users

Distributions

Messages per team

Features:

  • Horizontal or vertical bars

  • Hover tooltips

  • Legend for multiple series

Pie Chart

Best for showing proportions.

Use Case
Example

Distribution

Model usage breakdown

Composition

Sentiment distribution

Shares

Team contribution

Features:

  • Interactive segments

  • Percentage labels

  • Color-coded legend

Word Cloud

Best for text analysis.

Use Case
Example

Frequent topics

Common keywords

Trends

Popular terms

Patterns

User language analysis

Features:

  • Size = frequency

  • Color coding

  • Interactive hover

Line Chart

Best for trends over time.

Use Case
Example

Time series

Daily active users

Trends

Weekly satisfaction

Comparisons

Month-over-month

Features:

  • Multiple data series

  • Time axis

  • Trend indicators

Table

Best for detailed data.

Use Case
Example

Rankings

Top users list

Details

Recent conversations

Lists

Active prompts

Configuration:

  • 5 to 100 rows

  • Automatic sorting

  • Responsive columns


Card Sizing and Layout

Grid System

The dashboard uses a 4-column grid. Cards can span:

Size
Columns
Rows
Grid Space

Small

1

1

1x1

Medium

2

1

2x1

Large

3

2

3x2

Wide

4

1

4x1

Tall

2

2

2x2

Extra Large

4

2

4x2

Size Recommendations

Visualization
Recommended Size

Single metric

Small

Simple chart

Medium

Complex chart

Large or Tall

Table

Wide or Extra Large

Word cloud

Large or Tall

Layout Tips

  • Place important metrics at top left

  • Group related cards together

  • Use wide cards for tables

  • Balance visual weight across rows


Page Management

Default Pages

Page
Purpose

Overview

High-level metrics summary

User Tracking

User activity and engagement

Topic Analysis

Conversation topics and trends

Guidelines & Society

Compliance and policy metrics

Evaluators

Evaluation and quality scores

My Cards

Your custom created cards

Configure

Dashboard settings

Customizing Pages

Adding Cards to a Page

  1. Navigate to the target page

  2. Click Edit Page

  3. Drag cards from available list

  4. Position on grid

  5. Click Save Page

Removing Cards

  1. Click Edit Page

  2. Click the X on the card

  3. Card removed from page (not deleted)

  4. Click Save Page

Reordering Cards

  1. Click Edit Page

  2. Drag cards to new positions

  3. Grid snaps cards to positions

  4. Click Save Page


Team Filtering

Filter analytics data by team.

Using the Team Filter

  1. Click the Team dropdown in header

  2. Select a specific team

  3. Or select All Teams for aggregate

  4. All cards update automatically

Screenshot: Team Filter

Filter Behavior

Selection
Data Shown

Specific team

Only that team's data

All Teams

Aggregated across all

My Teams

Teams you belong to

Permissions

  • Only see teams you have access to

  • Enterprise admins see all teams

  • Team filter respects RBAC


Date Range Selection

Filter data by time period.

Using Date Picker

  1. Click the calendar icon in header

  2. Select start and end dates

  3. Or choose preset range

  4. All cards update automatically

Preset Ranges

Preset
Period

Today

Current day

Yesterday

Previous day

Last 7 Days

Past week

Last 30 Days

Past month

This Month

Current month

Last Month

Previous month

Custom

Your selection

Date Range Tips

  • Shorter ranges = faster loading

  • Compare same periods for trends

  • Consider data volume for range selection


Import and Export

Transfer dashboard configurations between accounts.

Exporting Dashboard

  1. Click Export Dashboard in header

  2. JSON file downloads automatically

  3. File contains:

    • Complete layout

    • All page configurations

    • Card settings

    • Visual arrangements

Filename: promptowl-dashboard-config-YYYY-MM-DD.json

Importing Dashboard

  1. Click Import Dashboard

  2. Select your JSON config file

  3. System validates and applies

  4. Layout imported immediately

Important: Share Analytics Prompts

After importing, you must share the underlying analytics prompts:

  1. Go to Monitor page

  2. Find the analytics prompts used by cards

  3. Share prompts with the target account

  4. Dashboard becomes functional

Transfer Workflow


Best Practices

Card Design

Do:

  • Use descriptive titles

  • Choose appropriate visualization

  • Size cards for content

  • Group related metrics

Don't:

  • Overcrowd pages

  • Use wrong chart types

  • Create duplicate cards

  • Ignore mobile view

Data Sources

Do:

  • Use focused analytics prompts

  • Generate structured output

  • Test before deploying

  • Document data sources

Don't:

  • Use overly complex prompts

  • Expect unstructured analysis

  • Ignore data quality

Performance

Do:

  • Limit cards per page (8-12 max)

  • Use appropriate date ranges

  • Choose efficient refresh rates

  • Archive unused cards

Don't:

  • Add too many cards

  • Use real-time for everything

  • Ignore loading times


Troubleshooting

"No data available"

Cause: No data in selected filters

Solutions:

  1. Expand date range

  2. Check team filter

  3. Verify data source has data

  4. Wait for data collection

"Data source not found"

Cause: Analytics prompt deleted or access revoked

Solutions:

  1. Check prompt still exists

  2. Verify your access permissions

  3. Re-select data source

  4. Contact prompt owner

"Processing error"

Cause: LLM analysis failed

Solutions:

  1. Check prompt configuration

  2. Verify API keys are valid

  3. Retry after a moment

  4. Review prompt output format

Cards Not Loading

Solutions:

  1. Refresh the page

  2. Check internet connection

  3. Clear browser cache

  4. Verify team/date filters

Import Failed

Solutions:

  1. Verify JSON file format

  2. Check file isn't corrupted

  3. Ensure file is valid export

  4. Try smaller configuration


Quick Reference

Visualization Selector

Data Type
Best Visualization

Single number

Number/Metric

Categories

Bar Chart

Proportions

Pie Chart

Time series

Line Chart

Text frequency

Word Cloud

Detailed list

Table

Size Quick Guide

Content
Recommended Size

KPI number

Small

Simple chart

Medium

Detailed chart

Large

Data table

Wide

Complex view

Extra Large

Refresh Intervals

Interval
Use Case

Real-time

Live dashboards

Hourly

Active monitoring

Daily

Standard reporting

Weekly

Summary metrics


Last updated