Skip to main content

How to Analyze Narratives for Users or Keywords

This article provides a highly effective prompt for extracting clear, distinct, and data-driven narratives from social-media conversations using XPOZ MCP.

M
Written by Mohar Bar
Updated over 3 weeks ago

Use coding tools and the XPOZ MCP and analyze the posts for {username, keywords} using the largest amount of posts that you can and extract 8–12 narratives based on RECURRING PHRASES and THEMES.

EXTRACTION GUIDELINES:

1. DETECT REPEATED VERBAL PATTERNS

- Exact or near-exact phrases appearing 3+ times

- Viral quotes, slogans, or specific accusations

- Lower frequency acceptable if engagement is high

2. PRESERVE DISTINCT TALKING POINTS

- Keep separate if different claims about the same subject

- Merge only if phrasing or meaning is truly identical

- Each narrative must be semantically unique

3. USE AUTHENTIC LANGUAGE

- Copy exact, sharp phrasing from posts

- Preserve original voice, tone, and emotional intensity

- Sound like the actual users, not analysts

- Avoid academic or overly neutral language

4. CREATE OPINIONATED NARRATIVE NAMES

- Reflect the posts' original language and tone

- Convey a clear stance or position

- Use direct, combative language where appropriate

- Sound passionate and engaged, not detached

5. ENSURE DISTINCTIVENESS

- Each narrative must have unique semantic meaning

- Maximize distinction between selected narratives

- Different angles on the same topic are allowed if the claims differ

6. CALCULATE POPULARITY

- Base popularity on phrase frequency in the dataset (and optionally engagement)

- Express as an integer percentage (0–100)

Return 8–12 MAXIMALLY DISTINCT narratives with:

- Name (10–120 chars)

- Description (max 300 chars)

- Popularity (integer, percentage of total discourse)

FINAL OUTPUT REQUIREMENT

Deliver a fully visual, beautifully designed HTML file, including:

- Responsive layout (desktop and mobile friendly)

- Colorful charts (SVG or embedded visuals) that illustrate narrative popularity, engagement, and other key metrics

- Clean typography and clear hierarchy (headings, subheadings, body text)

- Structured sections (e.g., Overview, Methodology, Narrative Breakdown, Key Metrics, Example Posts, Conclusions)

- Real posts with metrics (platform, date, engagement numbers, and relevance to each narrative)

- A professional look suitable for direct client delivery

The HTML must be complete and self-contained, ready to be saved as an .html file and opened in a browser.

Did this answer your question?