-
-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Milestone
Description
Goal
To provide teachers with real-time, actionable analytics within the context of a specific chat. This panel focuses on answering the question, "What are my students struggling with in this topic right now?"
To-Do List
- Create a Topic-Tagging Service:
- Implement a background service that takes a student's query and uses an LLM to tag it with the relevant lesson from the approved
Roadmap. This tag is stored with theSessionHistory.
- Implement a background service that takes a student's query and uses an LLM to tag it with the relevant lesson from the approved
- Create an Insight Generation Service:
- Create a new
AnalyticsServicemethod that analyzes the query tags for a specificchat_id. - This service will identify which roadmap topics are generating the most questions and create anonymized summaries of the most common questions asked.
- Create a new
- Implement the "Classroom Insights" UI:
- In the three-panel chat view, create a new tab in the right-hand sidebar (
Panel 3) labeled "Insights". - This panel will call the new analytics endpoint and display key findings, such as:
π₯ **Top Point of Confusion:** "Lesson 3: Special Relativity"π **Most Asked Questions about this topic:**"1. What is the twin paradox?""2. How does gravity affect time dilation?"
- In the three-panel chat view, create a new tab in the right-hand sidebar (
- Implement Action Buttons:
- Next to each insight, provide simple buttons for the teacher to take action, such as
[Generate a clarifying worksheet on this topic]or[Draft a note to all students who asked about this].
- Next to each insight, provide simple buttons for the teacher to take action, such as
Acceptance Criteria
- The system automatically identifies and flags topics where students are struggling within a specific chat.
- The teacher is presented with an anonymized summary of the most common questions, not the raw queries.
- The teacher can use one-click actions to generate support materials based on the AI's suggestions.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels