TrackMyEvents - An event tracking platform

What I built

I built an event tracking system which enables users to track the activity on their applications (web or mobile) using a simple API call. It provides the user with various analytics like daily active users over the last month, page-wise event data & many more.

Category Submission:

Smooth Shifters
Integration Innovators

App Link

https://track-my-events.vercel.app/
Login Credentials:
Email: [e-mail]
Password: testpassword

Screenshots

Dashboard

Profile

SDK Setup

Login

Register

Description

TrackMyEvents is a platform based on a microservices architecture which enables users to track the activity on their applications (web or mobile) using a simple API call. It provides the user with various analytics like daily active users over the last month, page-wise event data & many more.

The tracking system uses PostgreSQL as its primary database & Redis Pub/Sub for queuing events to store events in the database asynchronously. React is used for the frontend.

Link to Source Code

https://github.com/chirgjin/track-my-events

Permissive License

MIT

Background

Last year, I was looking for an event tracking solution which could provide me with high-level analytics for all the events along with the actual raw data which I can query at my convenience. However, I only found Google Analytics to be affordable and all other solutions were either not suited to my use case or very expensive. Due to this, I thought of making my own tracking solution as a test which proved to be very helpful to my learning.

How I built it

Originally, the TrackMyEvents was using RedisStack as its primary database for everything but seeing that there was no managed Redis offering on Linode, I decided to switch Redis to PostgreSQL and used Prisma ORM for interacting with it. This was a new learning experience for me as I had never used Prisma earlier in any of my projects. While using Prisma I found out that while it has many good features (like awesome type safety & auto-completion), Prisma doesn't support DISTINCT in aggregations. In the tracking service, I had to compute various analytics for the users & most of them required COUNT DISTINCT functionality which led me to write raw queries for the analytic APIs which was also very fun & a new experience for me.

When building this project, I utilized Linode for various purposes:

• Managed PostgreSQL: I used a separate managed PostgreSQL database provided by Linode for tracking & user service
• Linode: I used a 1GB Nanode to host my Redis server for pub/sub & a 2 GB Linode to host all the 3 microservices.

During all of my earlier projects I used to follow Digital Ocean's guide for the VM's initial server setup but this time I followed Linode's guide which was quite a nice surprise for me. I found out for the first time about the dpkg-reconfigure & hostnamectl commands as well as hardening the SSH Access.

Additional Resources/Info

Tech Stack

• Typescript + ExpressJs + Prisma ORM for backend
• Typescript + React.Js for frontend
• PostgreSQL for database
• Redis for Pub/Sub
• Vercel for frontend deployment

Conclusion

Overall, this was a very fun & challenging hackathon for me. I got to learn many new things from both Linode's documentation as well as Prisma's. The one thing I would definitely take care of is to avoid Prisma when I have to do analytical work next time 😝 and instead look for other alternatives. Another thing I'd like to mention to the Linode team is that while there's a category for EdgeWorkers in this hackathon, I couldn't find any information on EdgeWorkers on their website & even when I filled the form on Akamai's website, I didn't get any response from them.

Thanks to the Linode & Dev team for this opportunity. I am looking forward to hosting more of my projects on Linode 😁