I Was Tired of Sending My Users' Data to Google — So I Built My Own Analytics

I Was Tired of Sending My Users' Data to Google — So I Built My Own Analytics



Tópico: I Was Tired of Sending My Users' Data to Google — So I Built My Own Analytics
Categoria: Tutoriais | Programação & Tecnologia
Idioma Principal: Português (Conteúdo de Tecnologia)

Descrição do Conteúdo / Informações:
-------------------------------------------------------------------------
I'm a final year CS student. I build things in production.

Last year, I was working on [Ucoder](https://www.ucoder.in) — a full-stack platform I've been building since college. At some point I needed analytics. I wanted to know: are people actually using this? Which features are being used? Where are errors happening?

So I did what everyone does. I installed Google Analytics.

It worked. But something felt wrong.



The Problem


GA4 sends every user interaction to Google's servers. Every click, every session, every user's IP and device — all of it leaves your infrastructure and goes somewhere you don't control.

For a platform where I cared about user trust, that felt like a betrayal.

So I looked at alternatives:

•
Plausible — privacy-first, cookie-free, clean dashboard. But it only gives you pageviews, referrers, and bounce rate. No custom events, no error tracking, no real product analytics.

•
Mixpanel — powerful, but $28/month per 10K users. At scale, brutal. And the setup takes half a day just for event taxonomy.

•
Microsoft Clarity — session replays and heatmaps only. No event tracking.

•
Self-hosted Matomo — closer, but heavy to maintain and the UI feels dated.

None of them did what I actually needed:

Track real user behavior, capture errors, support custom events — without cookies, without sending data to a third party, and without a painful setup.

So I built it myself.



Introducing Ucoder Insights


Ucoder Insights is a privacy-first, cookie-free analytics platform with its own npm SDK.

It tracks:

• 📄 Page views (SPA-aware, works with Next.js, React, Vue)

• ⚡ Custom events

• 🔴 JavaScript errors (with stack traces)

• 📊 Core Web Vitals (LCP, FID, CLS)

• 🌐 Referrers, UTM params, device/browser info

All without cookies. All processed on your own terms.



How It Works — Technical Overview




The SDK


npm install ucoder-insight

import { initUcoderInsight } from 'ucoder-insight';

initUcoderInsight("projectid"{
// Your optional settings here
});

That's it. One function call. No configuration overhead.

The SDK is ~9.5KB gzipped. It uses a beacon API to send events so it never blocks the main thread. For SPAs, it listens to history.pushState and popstate to track client-side navigation automatically.



The Backend Architecture


The backend is built with Express 5 + TypeScript. When an event comes in:

• Request hits the Express API

• Passes through express-validator for input sanitisation

• Gets pushed into a BullMQ job queue backed by Redis

• A background worker processes the queue and writes to MongoDB

Why a queue? Because analytics should never slow down your app. If the ingestion endpoint is under load, events queue up and get processed asynchronously — your tracked application is never affected.

Client SDK
│
▼
Express API  ──→  BullMQ Queue (Redis)
│
▼
Worker Process
│
▼
MongoDB



Privacy by Design


•
No cookies — identification uses a daily-rotated hash of IP + User-Agent. No persistent tracking across days.

•
No PII storage — we never store  IPs or personal data.

•
Your data stays yours — events go to your project only, not a shared Google/Meta data warehouse.



What This Solves That Others Don't




1. Error Tracking + Analytics in One Place


With GA4 or Plausible, you need a separate tool (Sentry, LogRocket) for error tracking. With Ucoder Insights, JavaScript errors are captured automatically — with the page URL, user agent, and stack trace — right alongside your pageview and event data.

No context switching between dashboards.



2. Real Custom Events, Simply


import { trackEvent } from 'ucoder-insight';

// Track anything
trackEvent('signup_completed', { plan: 'free', method: 'email' });
trackEvent('feature_used', { feature: 'export_csv' });

Plausible's custom events require manual script tagging. Mixpanel requires you to plan an entire event taxonomy before you start. Here, you call one function.



3. SPA-First


Most analytics tools were built for traditional multi-page websites. Ucoder Insights was built with SPAs in mind from day one — it handles pushState navigation, React Router, Next.js App Router, and hash-based routing out of the box.



The Stack


Layer
Technology

SDK
TypeScript, Beacon API

Frontend Dashboard
Next.js 16, React 19, Tailwind CSS 4

Backend
Express 5, TypeScript, Socket.IO

Queue
BullMQ + Redis (ioredis)

Database
MongoDB (Mongoose 9)

Deployment
Docker + PM2 on Linux VPS



Current Status


The platform is live at insights.ucoder.in and running in production on Ucoder.in.

The npm package ucoder-insight is public at v1.1.3.

I'm a solo developer and this is a real project — not a side project I'll abandon. I use it myself, which means I have a real incentive to keep it working and improving.



What's Coming Next


•
Alert system — get notified on Slack/email when error rate spikes or traffic drops

•
Server-side proxy — bypass ad-blockers without compromising privacy

•
Funnel analysis — visualise conversion paths across custom events

•
API access — pull your analytics data into your own dashboards



Try It


npm install ucoder-insight

• 🔗 Platform: insights.ucoder.in

• 📦 npm: npmjs.com/package/ucoder-insight

• 💻 My portfolio: soumyadip.ucoder.in

If you're building something and tired of sending your users' data to Google — give it a try. And if you have feedback, I genuinely want to hear it.

Built by Soumyadip Maity — Full Stack Developer, final year CS student, West Bengal, India.