I Built a Sales Visualizer for a Real Business Problem (Quantium Software Engineering Simulation)

Source: Dev.to

End‑of‑Year Project: Quantium Software Engineering Simulation

“The best way to learn is by building something real.”

I wrapped up 2025 by completing Quantium’s Software Engineering simulation on Forage. After the Tata GenAI Data Analytics Challenge and several data‑driven apps, I was ready for another hands‑on project.

Want to try it yourself? Check out the simulation here (spoiler alert ahead!).

---

The Scenario – Software Engineer at Quantium

You play a software engineer in Quantium’s financial‑services business unit, working for the client Soul Foods.

This wasn’t a toy tutorial – it was a real‑world business question that needed a code‑driven solution.

The Challenge – Six Progressive Tasks

The simulation scaffolds the work so each task naturally builds on the previous one, just like a real project.

Task 1 – Set Up Local Development Environment

• Fork the repo.
• Create a Python virtual environment.
• Install dependencies (dash, pandas, etc.).

Takeaway: A clean, well‑organized workbench saves countless hours later.

Task 2 – Data Processing: The Art of Reshaping Data

Three messy CSVs contain transaction data for the entire morsel product line. I turned them into a tidy dataset.

Pipeline steps

1. Filter – Keep only rows for Pink Morsels.
2. Calculate – Compute Sales = quantity × price.
3. Normalize – Strip currency symbols, parse dates, standardise region names.
4. Output – Write a clean CSV with columns Sales, Date, Region.

I wrote a defensive ETL script with a flexible column‑detection helper (find_column) to cope with inconsistent naming.

Task 3 – Create the Dash Application

Built a minimal Dash app that:

• Shows a clear header stating the business question.
• Plots a line chart of daily sales.
• Adds a vertical marker on 2021‑01‑15 (price‑increase date).

The chart instantly answered Soul Foods’s question – you can see the impact.

Task 4 – Make It Interactive & Beautiful

Added features requested by the client:

• Radio buttons to filter by region (North, East, South, West, All).
• Custom CSS for a modern, clean look.
• Responsive layout for different screen sizes.

Dash callbacks make the chart update instantly when a region is selected.

Task 5 – Write a Test Suite

Production‑grade code needs tests. Using pytest together with Dash’s testing utilities, I verified:

• Header presence.
• Graph rendering.
• Region picker functionality.

Recursive component finders traverse the layout tree, protecting against regressions as the code evolves.

Task 6 – Automate Everything with CI

A Bash script ties everything together for continuous integration:

#!/usr/bin/env bash
set -euo pipefail

# Activate virtual environment
source .venv/bin/activate

# Install dependencies (if needed)
pip install -r requirements.txt

# Run the full test suite
pytest

The script returns proper exit codes, enabling CI engines to report success/failure automatically.

Why This Challenge Is Cool

What I Built

Tech Stack

• Python 3.9
• Pandas – data wrangling
• Dash & Plotly – interactive web app & visualizations
• Pytest – unit & integration testing
• Bash – CI automation script

Final Thoughts

The Quantium simulation gave me a compact, end‑to‑end experience that mirrors a real software‑engineering project. I sharpened my data‑processing, visualization, testing, and CI skills while delivering a tangible business insight. If you enjoy building practical solutions, I highly recommend giving the simulation a try!

The Foundation

• Dash – Web framework for data applications
• Plotly Express – Interactive, beautiful charts
• Pandas – Data‑manipulation powerhouse
• Pytest – Testing framework
• Bash – CI automation scripting
• CSS – Custom styling for a polished UI

👉 My Submitted Repo
👉 Original Source Code

Key Takeaways

• Start with clean data – Garbage in, garbage out. Invest in robust ETL.
• Let the data speak – Simple visualisations often tell better stories than complex ones.
• Build for humans – A pretty UI isn’t vanity; it’s usability.
• Test early, test often – Even simple tests catch real bugs.
• Automate the boring stuff – CI scripts save hours of manual work.
• Modular architecture wins – Separating data, viz, and web layers made iteration easy.

Try It Yourself

Then come back and tell me:

• How did you style your visualiser?
• What patterns did you discover in the data?
• Did the sales actually go up or down after the price increase? 😏

Potential Next Steps

Final Reflections

This project stretched me across roles: data engineer, frontend developer, and DevOps practitioner. Real software problems don’t come in neat boxes. I walked away with a working application, clean architecture, and practical experience with a modern data‑visualisation stack.

The answer to Soul Foods’s question? Run the app yourself and find out. The data doesn’t lie. 📊