← All projects
ContactResume
Project

Post-Conversation Analysis

Django REST API + Celery for support chats

Analyses customer support conversations across seven NLP metrics, with scheduled background processing and an admin dashboard.

PythonDjango RESTCeleryTextBlobNLTKSQLite
scroll
Overview

What it is

A Django REST API that automatically analyses customer support conversations across seven NLP metrics: sentiment, empathy, clarity, relevance, accuracy, completeness and response time. Celery runs scheduled background analysis and a Django admin dashboard is used to review results. It started as my HackerHub8 internship project.

Upload conversations

POST a user/AI message sequence and all content is stored.

Seven metrics

Every conversation is scored on clarity, relevance, accuracy, completeness, empathy, sentiment and average response time, plus fallback detection, resolution status and escalation flags.

On-demand or scheduled

Trigger analysis through the API, or let Celery Beat process pending conversations every minute.

Reports API

List and filter analyses by sentiment and resolution, ordered by overall score.

Admin dashboard

Manage conversations, messages and analyses from Django admin.

Tested

A 32-test suite covering the analyzers, API, filters and tasks, with CI.

Flow

How it works

Upload

Send a conversation as a sequence of user and AI messages.

Analyse

Pure-function analyzers score the seven metrics, via the API or the Celery schedule.

Review

Filter reports by sentiment and resolution, or browse in the admin dashboard.

SQLite by default (PostgreSQL via settings). Scheduled analysis needs Redis for the Celery broker.

Run it

git clone https://github.com/Prajit-22/post-conversation-analysis.git cd post-conversation-analysis pip install -r requirements.txt python manage.py migrate python manage.py runserver
Next projectSecond Brain →