Skip to main content

​
Overview

The --analyze flag activates Grok’s code_interpreter tool, enabling quantitative analysis of X search results. Grok writes and executes code to compute:
  • Engagement statistics (likes, reposts, replies)
  • Posting frequency and velocity
  • Trend patterns over time
  • Sentiment distribution
  • Geographic signals
  • Account influence metrics
This transforms raw X posts into actionable data insights.

​
How It Works

Add the --analyze flag (or -a) to any X Search query: 
/grok-x "<query>" --analyze
Grok will:
  1. Use x_search to gather relevant posts
  2. Activate code_interpreter to process the data
  3. Generate statistics, tables, and trend analysis
  4. Present findings with specific numbers and patterns
The --analyze flag cannot be combined with --extract. Choose quantitative analysis or structured JSON output, but not both.

​
What It Can Analyze

  • Engagement metrics: Like counts, repost ratios, reply volumes
  • Posting patterns: Daily/hourly frequency, peak activity times
  • Trend velocity: Rate of growth or decline in discussion volume
  • Sentiment distribution: Breakdown of positive/negative/neutral posts
  • Company/topic mentions: Frequency tables by entity
  • Geographic distribution: Where posts are originating from
  • Account influence: Amplification patterns and reach

​
Real Example: Tech Layoffs Analysis

/grok-x "Tech layoff announcements this month" --analyze --from 2026-02-01
Output:
Tech Layoff Activity — February 2026 Analysis Grok searched X for layoff announcements and ran quantitative analysis via code interpreter across 847 relevant posts from Feb 1–Mar 4, 2026. By company (posts mentioning layoffs):
CompanyPostsEstimated headcount reduction
Salesforce143~1,200
Intel98~2,800
Snap67~400
Twilio54~350
Workday41~600
Posting velocity: Layoff-related posts peaked on Feb 14 (Valentine’s Day — 203 posts in 24h) and again on Feb 28 (end-of-month HR cycles — 178 posts). Sentiment in layoff posts: 61% negative/critical, 24% neutral/informational, 15% supportive of company rationale. The supportive cluster skews toward accounts in finance and investor relations. Geographic signal: 71% of geo-tagged posts originated from US accounts, 14% from India (Salesforce and Workday have large India engineering hubs), 8% from UK. Most amplified single post: An Intel engineer’s firsthand account of receiving a severance email with 48 hours notice — 94K likes, 41K reposts, 6.2K replies. Conclusion: February 2026 is tracking ~18% above the rolling 6-month average for layoff post volume. The Intel and Salesforce announcements are dominating signal. No signs of a sectoral recovery narrative emerging yet.

​
Data Tables and Insights

Grok’s code_interpreter produces:

​
Frequency Tables

| Entity | Post Count | Percentage |
|--------|------------|------------|
| Value1 | 143        | 34.2%      |
| Value2 | 98         | 23.4%      |

​
Time Series Analysis

  • Peak posting times
  • Day-over-day trends
  • Weekly/monthly patterns

​
Engagement Metrics

  • Average likes per post
  • Repost-to-like ratios
  • Viral content identification
  • Amplification patterns

​
Statistical Summaries

  • Means, medians, and distributions
  • Comparative analysis (current vs. baseline)
  • Outlier detection

​
Example Use Cases

​
Engagement trend analysis

/grok-x "@elonmusk engagement this month" --analyze --handles elonmusk --from 2026-02-01
Analyzes posting frequency, engagement rates, and content themes for a specific account.

​
Topic velocity tracking

/grok-x "artificial general intelligence discussions" --analyze --from 2026-01-01
Tracks how discussion volume has changed over time and identifies peak periods.

​
Comparative sentiment

/grok-x "Bitcoin vs Ethereum sentiment" --analyze
Compares discussion volume, sentiment distribution, and engagement patterns between two topics.

​
Event impact measurement

/grok-x "Product launch reactions" --analyze --from 2026-02-20 --to 2026-02-27
Quantifies reaction patterns in the week following a specific event.

​
Combining with Other Flags

The --analyze flag works with:
  • --handles / --exclude: Focus analysis on specific accounts
  • --from / --to: Analyze specific time periods
  • --images / --videos: Include media-rich posts in analysis
  • --citations: Get source URLs for verification

​
Example: Video content analysis

/grok-x "Tesla discussion" --analyze --videos --from 2026-02-01
Analyzes posts containing video content about Tesla, including frequency and engagement patterns.
Constraint: --analyze and --extract cannot be combined. Use --analyze for quantitative insights or --extract for structured JSON, but not both in the same query.

​
When to Use Data Analysis

Choose --analyze when you need:
  • Numbers and statistics rather than narrative summaries
  • Trend identification across time periods
  • Comparative analysis between entities or topics
  • Engagement metrics to measure amplification
  • Pattern detection in posting behavior or sentiment
For structured, typed JSON output instead, use --extract.

Build docs developers (and LLMs) love

Get started for freeTalk to us