Data Explorer

The Data Explorer is an interactive visual tool for exploratory data analysis, feature engineering, and data reporting of your chatbot’s conversations.

Configuration

• View - Data Explorer allows you to select a subset of your conversational data for viewing or analysis. These 5 parameters are different ways of aggregating subsets of your JSON payload to explore your data further.
  • Messages - Count of unique messages
  • Users - Count of unique users across all messages
  • Intents - Count of all unique incoming/outgoing intents
  • General Field - Specific fields associated with parameters related to our platform. You can select to view from any of the following general fields and apply one of the following aggregate functions: count, count_distinct, sum, avg, min, and max
    • platformUserID - the ID for your chatbot’s platform
    • sessionID - the ID for a distinct conversation session
    • Sentiment - the corresponding sentiment value for a session
    • Datetime - the timestamp for each message
  • JSON Field - Specific fields associated with the JSON payload. This allows for specific/custom fields that you pass through your JSON payload (e.g. intentConfidence). You can select to view any specific category from your JSON payload and apply one of the following aggregate functions: count, count_distinct, sum, and avg.
    • Below is an example of selecting to view the Intent Confidence from the JSON metadata
    • Example - fm.metadata/intentConfidence count as IntentConfidence
• Filter - An initial filter of your bot’s data before viewing the output. A filter selects a subset of your bot’s data determined by specific constraints. You can find the list of possible filter constraints in the Glossary section below.
• Breakdown - A breakdown classifies the resulting data into segments based on the selected category. You can find the list of possible breakdowns in the Glossary section below.

Output

• Date Range - You can select the time range for the conversational history of your chatbot
• Linear vs Logarithmic - You can select to view the time series graph as either linear or logarithmic
• Output/Graph Type - The following are the different graph views available for Data Explorer:
  • Line
  • Stacked Line
  • Bar
  • Stacked Bar
  • Table
  • Condensed Table
  • Pie
  • KPI

Settings

• Data Explorer let’s you save and download your unique DE configurations.
  • Load - load a previously saved DE configuration
  • Save - saves the current configuration
  • Save As New - saves as a new configuration
  • New - resets the configuration (pretty much a refresh)
  • Download CSV - To download as a CSV file click the ellipses … in the upper right hand corner of the report

Use Cases

The following are a few popular use cases to explain how you might use Data Explorer to answer common questions regarding your data:

Demo video on using the Data Explorer to identify which intents and messages types are being sent by your users

• Use Case 1: View Breakdown of First Message based on Message Type & Intent
  • View: Messages
  • Filter: First Message In Session
  • Breakdown: Incoming, Message Type, Intent
  • Be able to identify which intents and message types are the first to be sent to your chatbot by users
• Use Case 2: View Intent Matching based on Keywords
  • View: Messages or Users
  • Filter: Text containing “keyword”
  • Breakdown: Intent, Text
  • Be able to easily identify Mishandled/Unhandled intents that should be Handled correctly
  • Example: Filtering for all messages containing the keyword “Help”
    • Identify which “help” messages are mapping to incorrect and unhandled intents to use as training phrases
• Use Case 3: View Average Intent Confidence for Incoming Messages
  • View:
    • Messages
    • Users
    • sessionId - count_distinct
    • fm.metadata/intentConfidence - avg
  • Filter: Incoming
  • Breakdown: Intent
  • Be able to identify which intents where most often triggered and the respective users, sessions, and intent confidence score

Glossary