Nov 27, 2024

Hey Claude, help me analyze Bluesky data.

This is the full, unedited transcript of our conversation with Claude, whose context-awareness is provided by a v0 Tinybird MCP Server.
Claude
AI Assistant
Tinybird MCP Server
I provide context

Below is the full transcript of a conversation with Claude, made context-aware using an early implementation of a Model Context Protocol (MCP) Server for Tinybird Workspaces. The Tinybird MCP Server v0 accepts arguments for a Tinybird host and token, so we supplied the credentials for a Tinybird Workspace capturing thousands of events per second from the Bluesky Firehose. Here's what Claude told us.


Hey there! I see you've chosen to explore Bluesky data. Let's dive in and discover some interesting insights!

Bluesky is a decentralized social media platform that's been gaining attention. It offers a fascinating dataset that could reveal trends in user engagement, content sharing patterns, and community growth. Let me take a quick look at what data we have available to explore.

Inspecting the data...

View Result from list-data-sources from tinybird_mcp_claude (local)

View Result from list-pipes from tinybird_mcp_claude (local)

  • Follow activity tracking (per second and minute)
  • Daily account creation and deletion metrics
  • Post counts by user
  • Raw event data

Let me show you some initial insights we could explore. Would you like to:

A) Analyze user growth trends by looking at daily account creations vs deletions

B) Examine user engagement through posting behavior

C) Investigate follow activity patterns to understand network growth

D) Look at overall platform activity trends

Which aspect would you like to explore first? Each of these will help us understand different aspects of Bluesky's ecosystem and user behavior.

I've started our insights memo with an overview of the available data. As we explore further, we'll add more detailed insights to build a comprehensive understanding of Bluesky's platform dynamics. What would you like to explore first?

C

Let's dive into the follow activity patterns! This will give us great insights into how the Bluesky network is growing and how users are connecting with each other.

I'll analyze the follows per minute data to get a good view of recent trends.

View Result from run-select-query from tinybird_mcp_claude (local)

Interesting! Let me get a broader view of the pattern throughout the day.

View Result from run-select-query from tinybird_mcp_claude (local)

Fascinating! The data shows very active network growth with around 20,000 new follows happening every minute! We're seeing peaks of over 1.2 million follows per hour, which indicates extremely healthy user engagement and community building.

Would you like to:

A) Dig deeper into peak activity times to understand when users are most engaged

B) Compare follow activity with new account creation to see how new users contribute to network growth

C) Look at the relationship between posting activity and follow patterns

D) Analyze the distribution of follows to understand if growth is evenly spread or concentrated

What interests you most? Each option will give us different insights into how the Bluesky community is growing and connecting.

I've added our follow activity insights to the memo, which is being updated in real-time. You can find these insights by clicking the paperclip icon, selecting the MCP menu (two electrical plugs connecting), and choosing "Insights Memo".

C

Let's explore the relationship between posting activity and follow patterns! This will help us understand how content creation relates to network growth.

First, let me get data about posting activity:

View Result from run-select-query from tinybird_mcp_claude (local)

Now, let me look at the distribution of posting activity:

View Result from run-select-query from tinybird_mcp_claude (local)

Let me try a simpler query to understand posting distribution:

View Result from run-select-query from tinybird_mcp_claude (local)

Let me try the pre-built pipe endpoint for average posts:

View Result from request-pipe-data from tinybird_mcp_claude (local)

Interesting findings! Looking at the relationship between posting and follow activity, we can see:

  • There are about 1.15 million unique users who have posted content
  • These users have created approximately 7.7 million posts total
  • The average user has made about 7 posts
  • This substantial content creation appears to drive the high follow rates we saw earlier (around 20,000 follows per minute)

Would you like to:

A) Analyze how posting frequency correlates with follower growth

B) Look at daily account creation trends to see how new users contribute to content creation

C) Investigate peak posting times compared to peak follow times

D) Explore the ratio of posts to follows over time

Which aspect would you like to explore next? I've added our latest insights about content creation and its relationship to network growth to the memo. You can access these updated insights through the paperclip icon → MCP menu → "Insights Memo".

A

Let's analyze how posting frequency relates to follower growth! This will give us valuable insights into how content creation drives network expansion.

View Result from run-select-query from tinybird_mcp_claude (local)

Let me check the post activity over the same period:

View Result from run-select-query from tinybird_mcp_claude (local)

Let me try looking at recent post distribution:

View Result from run-select-query from tinybird_mcp_claude (local)

Fascinating findings about the relationship between posting frequency and network growth! We've discovered:

  • There are over 24,000 highly active creators who have posted 50 or more times
  • These active creators average about 120 posts each
  • This concentrated content creation appears to drive significant follow activity, with recent data showing over 2.3 million follows in a single day
  • The platform shows a healthy mix of casual and power users, with power users driving substantial engagement

Would you like to:

A) Deep dive into the characteristics of these top content creators
B) Analyze the daily patterns of follows to see when engagement peaks
C) Investigate how new user follows differ from established user follows
D) Explore the geographic distribution of high-engagement users

Which aspect interests you most? I've added our latest insights about the correlation between posting frequency and network growth to the memo. You can access the updated insights through the paperclip icon → MCP menu → "Insights Memo".

Let's build a comprehensive picture of how content drives growth on Bluesky!

wow such insights, much interesting - can you turn this into a dashboard for me?

I'll create a comprehensive dashboard that visualizes our key findings about Bluesky's platform dynamics!

I've created a comprehensive dashboard that visualizes all our key findings! The dashboard includes:

  • A real-time follows per hour chart showing the network growth patterns
  • Platform engagement metrics displaying total posts, unique users, and active creators
  • Key performance indicators showing:
  • Average posts per user (7)
  • Follows per minute (20K)
  • User distribution breakdown showing the split between power users, regular users, and casual users

The dashboard uses a clean, modern design with interactive charts that allow you to hover for detailed information. The color scheme uses different shades of blue to represent varying levels of engagement.

Would you like me to:

A) Add more specific metrics about content creation patterns?

B) Include geographic distribution data?

C) Add trend analysis over longer time periods?

D) Create more detailed breakdowns of user categories?

I'm continuing to update our insights memo, which you can access through the paperclip icon → MCP menu → "Insights Memo". The dashboard complements these insights by providing visual representations of our key findings.

editor note: Here is the dashboard that Claude built.

More Resources:

Do you like this post?

Related posts

More Data, More Apps: Improving data ingestion in Tinybird
Tinybird is out of beta and open to everyone
How I replaced Google Analytics with Retool and Tinybird, Part 1
Building an enterprise-grade real-time analytics platform
Behind the scenes of Tinybird's big frontend refactor
What it takes to build a real-time recommendation system
Migrating from Rockset? See how Tinybird features compare
A new dashboard for Tinybird Analytics
Low-code analytics with James Devonport of UserLoop

Tinybird

Team

May 05, 2023
Operational Analytics in Real Time with Tinybird and Retool

Build fast data products, faster.

Try Tinybird and bring your data sources together and enable engineers to build with data in minutes. No credit card required, free to get started.
Need more? Contact sales for Enterprise support.