NirajanKhadka/canadian-job-market-analytics

This is a well-scoped portfolio analytics project with a clear story: it takes 20,462 Canadian data job postings, normalizes them into PostgreSQL, and turns the results into a deployed Streamlit dashboard with practical views for skills, salaries, cities, and posting trends. The README does a good job making the project approachable. The screenshots, data source table, PostgreSQL table summary, and live Streamlit link give readers a quick sense of what was built and why it matters. I also like that the repo separates the dashboard from the pipeline and notebook work instead of leaving everything in one analysis file; dashboard/app.py, the dashboard/pages/ modules, and the pipeline/ scripts make the project feel more like a real data product than a one-off notebook.

The strongest part is that the project connects analysis to user-facing output. The dashboard is not just showing static charts; it has filters for role, city, salary, and remote work, and the SQL-backed KPI cards make the app feel grounded in the underlying dataset. The EDA highlights are useful too, especially the salary coverage caveat. Calling out that only 4.6% of postings include salary makes the salary findings more credible because it shows you are aware of data limitations rather than overclaiming.

The main thing I would address before promoting this more widely is repository hygiene and trust. There is an open issue reporting a possible PostgreSQL credential leak, so rotating any affected credentials and confirming secrets are fully removed from git history should be treated as urgent. I’d also add the actual LICENSE file since the README says MIT but GitHub does not detect a license. After that, adding a short schema diagram, a “how to reproduce the pipeline” section with the exact script order, and a small test or validation script for row counts/null handling would make the project easier for other analysts to run confidently. Overall, this is a useful, concrete data analytics project with a nice deployed surface; tightening security, licensing, and reproducibility would make it much stronger.