Hi everyone π
Iβm Vinayak, a Java backend developer from Pune, India
and a member of the NVIDIA AI Aerial / 6G Developer Program.
Over the past few weeks Iβve been building something I
couldnβt find anywhere β a visual, interactive, shareable
web dashboard on top of NVIDIA Sionna simulation results.
Iβm calling it Sionna Visualizer.
π Live Demo: https://sionna-visualizer.vercel.app
π GitHub: GitHub - Vinu2111/sionna-visualizer: A full-stack web dashboard for NVIDIA Sionna 6G simulation results. Angular + Java + Python + PostgreSQL. Β· GitHub
ββββββββββββββββββββββββββββββββββββββββ
THE PROBLEM I KEPT RUNNING INTO
ββββββββββββββββββββββββββββββββββββββββ
Sionna outputs raw Python data β tensors, numbers,
Jupyter notebook graphs. Every researcher Iβve spoken
to has the same workflow: run simulation, copy data
to Excel or matplotlib, generate chart manually,
share a screenshot.
There was no web-based visual layer. No shareable links.
No history. No comparison tool. I built one.
ββββββββββββββββββββββββββββββββββββββββ
WHAT IT DOES RIGHT NOW
ββββββββββββββββββββββββββββββββββββββββ
Three simulation types β all running real mathematics:
-
BER vs SNR Simulation
β Real AWGN channel model
β Two curves: theoretical (closed-form erfc equations)- simulated (Monte Carlo, 100,000 bits)
β BPSK / QPSK / 16QAM / 64QAM
β Logarithmic chart, 25 SNR points
β 94ms on CPU β no GPU required
- simulated (Monte Carlo, 100,000 bits)
-
Beam Pattern Visualization
β Uniform Linear Array (ULA) model
β 8 / 16 / 32 / 64 antenna elements
β 28 GHz / 39 GHz / 60 GHz / 77 GHz
β Polar chart showing main lobe + side lobes
β 16-element ULA gives -13.15 dB SLL β textbook correct
β Steering angle control: -90Β° to +90Β° -
Modulation Comparison
β BPSK vs QPSK vs 16QAM vs 64QAM simultaneously
β Crossover points computed automatically
β Spectral efficiency tradeoff visible instantly
Plus:
β Every simulation saved to PostgreSQL with full history
β Shareable public link for every result
β Simulation comparison tool β select any two from
history, overlay on one chart, winner computed automatically
β JWT authentication
β Full mobile responsive layout
ββββββββββββββββββββββββββββββββββββββββ
TECH STACK
ββββββββββββββββββββββββββββββββββββββββ
Frontend: Angular 17 + Chart.js (Vercel)
Backend: Java 17 + Spring Boot 3 (Railway)
Bridge: Python 3.10 + FastAPI + Sionna (Railway)
Database: PostgreSQL (Railway)
Auth: Spring Security + JWT
Fully open source under MIT License.
docker-compose up runs the entire stack locally.
ββββββββββββββββββββββββββββββββββββββββ
WHAT IS NOT BUILT YET
ββββββββββββββββββββββββββββββββββββββββ
Being transparent about current limitations:
β GPU Sionna simulation β currently using CPU-compatible
AWGN model. Full GPU Sionna integration is next.
β AODT (city-scale digital twin) integration β planned
β MIMO channel visualization β planned
β Ray tracing simulation support β planned
ββββββββββββββββββββββββββββββββββββββββ
WHY IβM POSTING HERE
ββββββββββββββββββββββββββββββββββββββββ
This community is exactly who I built this for.
I have three genuine questions for anyone who uses
Sionna in their research:
-
What Sionna simulation outputs do you wish you
could visualize that arenβt in the tool yet? -
What would make this actually useful in your
daily research workflow? -
Would a REST API that lets you POST your own
Sionna output JSON and get back a chart URL
be useful?
Any feedback β even harsh β is genuinely appreciated.
This is being built in public and your input directly
shapes what gets built next.
Full build-in-public journey documented at:
vinayak6g.hashnode.dev
Thank you for your time.
β Vinayak