Senthilsweb
Search
Analytics · DuckDB-WASM

What the writing actually reached

Every number here comes from my own LinkedIn export, converted to parquet and queried in your browser. The charts render server-side; the engine loads afterwards, only to power the controls. Each tile will show you the SQL behind it.

Aug 3, 2025 — Aug 2, 2026·4 parquet files
Period
Metric
engine idle — charts are server-rendered
49,762
Impressions
22,880
Members reached
1,051
Followers
52
Posts measured
SQL
SELECT impressions, members_reached, total_followers
FROM overview
ORDER BY snapshot_date DESC LIMIT 1;

Daily impressions

7-day average
Jul 28, 2025peak 3,131Aug 2, 2026
SQL
SELECT day, impressions, engagements, new_followers
FROM daily
WHERE day >= ?          -- period selector
ORDER BY day;

By month

impressions
Jul '25
Aug
Sep
Oct
Nov
Dec
Jan '26
Feb
Mar
Apr
May
Jun
Jul
Aug
SQL
SELECT date_trunc('month', day) AS month,
       sum(impressions)        AS impressions
FROM daily
WHERE day >= ?
GROUP BY 1 ORDER BY 1;

By country

followers, from top metros
  • India37%
  • United States16%
  • United Kingdom2%
  • Singapore2%
  • Not disclosed43%

LinkedIn breaks out only the top metro areas, so 57% of the audience can be attributed to a country. The rest is not reported.

SQL
-- LinkedIn gives metro areas, not countries. The country
-- column is mapped on import, so the gap stays visible.
SELECT country, sum(percentage) AS pct
FROM audience
WHERE demographic_kind = 'audience'
  AND category = 'Location' AND country <> ''
GROUP BY 1 ORDER BY pct DESC;

Top 10 posts

by impressions
#PostImpressionsEngagementsRate
1I Left My Job Hunt Running — Now It's a Public Dataset Anyone Can QueryJul 15, 2026 · article · 10 eng · 0.2%4,350100.2%
2One Core, Four Doors — Shipping the Same AI Agent as CLI, REST, MCP Tool, and ChatJul 12, 2026 · article · 16 eng · 0.4%4,017160.4%
3Telemetry and Observability for Claude CodeJul 7, 2026 · article · 40 eng · 1.1%3,646401.1%
4Job Hunting Is a Data Problem. I Treated It Like One.Jul 3, 2026 · article · 45 eng · 1.5%2,904451.5%
5Specs, Not Vibes — Building a Production Agent with AI-DLCJul 14, 2026 · article · 4 eng · 0.2%2,59140.2%
6homelab hybridcloud handsonengineeringJul 12, 2026 · 26 eng · 1.4%1,804261.4%
7Introducing Claimwise — End-to-End Data Lakehouse Storytelling with dbt, Databricks/Snowflake, and Unity Catalog, From Raw Data to DashboardJul 18, 2026 · article · 15 eng · 0.9%1,755150.9%
8My View: The AI Era Is Distilling and Rewarding Elegant EngineersJul 13, 2026 · article · 11 eng · 0.7%1,682110.7%
9Looking Back on 25 Years in Software DevelopmentDec 31, 2025 · article · 43 eng · 2.6%1,680432.6%
10OpenLineage Viewer: Databricks-Style Lineage for Local dbt, in Pure HTMLJul 20, 2026 · article · 11 eng · 0.8%1,448110.8%
SQL
SELECT coalesce(nullif(matched_title,''), topic_slug) AS title,
       impressions, engagements,
       round(engagements * 100.0 / nullif(impressions,0), 2) AS rate
FROM top_posts
ORDER BY impressions DESC LIMIT 10;

Audience

share of followers

Seniority

  • Senior44%
  • Entry13%
  • Director11%
  • Manager7%
  • VP6%
  • CXO5%
  • Owner3%
  • Partner2%

Industry

  • IT Services and IT Consulting36%
  • Software Development21%
  • Financial Services4%
  • Human Resources Services3%
  • Staffing and Recruiting3%
  • Technology, Information and Internet3%
  • Business Consulting and Services2%
  • Banking1%
  • Hospitals and Health Care1%
  • Semiconductor Manufacturing1%

Job title

  • Software Engineer6%
  • Founder3%
  • Chief Executive Officer2%
  • Co-Founder2%
  • Project Manager2%
  • Solutions Architect2%
  • Data Engineer1%
  • Professor1%
  • Architect< 1%

Company

  • Pentaho3%
  • Hitachi Vantara2%
  • HCLTech1%
  • IBM1%
  • Tata Consultancy Services1%
  • Accenture< 1%
  • Cognizant< 1%
  • Infosys< 1%
  • Vmoksha Technologies Pvt. Ltd< 1%

Company size

  • 10,001+ employees29%
  • 1,001-5,000 employees9%
  • 51-200 employees9%
  • 11-50 employees8%
  • 2-10 employees8%
  • 5,001-10,000 employees8%
  • 201-500 employees7%
  • 501-1,000 employees5%
  • 0-1 employees1%
SQL
SELECT category, value, percentage
FROM audience
WHERE demographic_kind = 'audience'
  AND snapshot_date = (SELECT max(snapshot_date) FROM audience)
ORDER BY category, percentage DESC;

Who saw the posts

share of viewers

LinkedIn reports this separately from the follower breakdown above: these are the people a post actually reached, most of whom do not follow him. It is the wider, less self-selected group of the two.

Seniority

  • Senior46%
  • Entry22%
  • Manager8%
  • Director6%
  • VP2%
  • CXO1%
  • Owner1%
  • Training1%
  • Partner< 1%

Industry

  • IT Services and IT Consulting41%
  • Software Development19%
  • Technology, Information and Internet5%
  • Financial Services4%
  • Banking2%
  • Data Infrastructure and Analytics2%
  • Hospitals and Health Care2%
  • Business Consulting and Services1%
  • Telecommunications1%

Job title

  • Software Engineer12%
  • Data Engineer6%
  • Solutions Architect2%
  • Architect1%
  • Data Architect1%
  • Data Scientist1%
  • Full Stack Engineer1%
  • Project Manager1%

Company

  • Tata Consultancy Services2%
  • Accenture1%
  • Cognizant1%
  • Deloitte< 1%
  • IBM< 1%
  • Infosys< 1%
  • Microsoft< 1%
  • Oracle< 1%

Company size

  • 10,001+ employees40%
  • 1,001-5,000 employees12%
  • 5,001-10,000 employees7%
  • 51-200 employees7%
  • 11-50 employees6%
  • 201-500 employees6%
  • 501-1,000 employees5%
  • 2-10 employees3%

Location

  • Greater Bengaluru Area10%
  • Greater Hyderabad Area5%
  • New York City Metropolitan Area4%
  • Dallas-Fort Worth Metroplex3%
  • Greater Chennai Area3%
  • Pune/Pimpri-Chinchwad Area3%
  • San Francisco Bay Area3%
  • Washington DC-Baltimore Area3%
  • Greater Delhi Area2%
SQL
-- Same table, the other population: people who saw a post,
-- whether or not they follow him.
SELECT category, value, percentage
FROM audience
WHERE demographic_kind = 'content'
  AND snapshot_date = (SELECT max(snapshot_date) FROM audience)
ORDER BY category, percentage DESC;

Source: LinkedIn aggregate analytics export, snapshot 2026-08-02. Imported with scripts/import-linkedin-analytics.mjs and published as CSV and parquet. Engagement figures are LinkedIn's own; nothing here is modelled or estimated.