Try it: https://drimranbhatti.github.io/global-firm-financials/
Inspiration
Small and mid-sized companies set payment terms, inventory targets and debt levels without knowing what is normal for their industry and country. Listed-company data answers that, but it sits behind expensive licences and needs a data team to use.
What it does
You pick a market (49 countries and regions) and an industry (20), enter your own ratios, and see your percentile among listed peers for receivable, inventory and payable days, cash conversion cycle, gross margin, operating ROA, debt, cash, current ratio, interest cover and sales growth. Each metric shows the peer median, the middle half of peers and where you sit, with a plain-English verdict and a copyable summary for a board pack or a loan file.
How I built it
- Built a panel of 607,799 annual statements from 52,792 listed non-financial companies in 123 countries (Compustat Global and North America, 2000-2020) in Python: de-duplication across databases, USD conversion with World Bank rates, winsorised ratios.
- Computed decile tables for fiscal 2017-2019 by market and industry, keeping only cells with at least 30 firms so that no number describes a single company.
- Wrote a single self-contained HTML/JavaScript page: no server, no tracking, and nothing the user types leaves the browser. Hosted free on GitHub Pages.
Challenges I ran into
Keeping the licensed data private while still publishing something useful: the tool ships only aggregate percentiles, and the build script that produces them is public so anyone with the licence can reproduce it.
Accomplishments that I'm proud of
- 49 markets and 480 market-industry cells.
- Part of a 14-notebook research repository that also includes an ML distress model (AUC 0.83 vs 0.59 for Altman Z''), a Beneish M-score study and a worldwide Shariah screen.
What I learned
CFOs and lenders trust a benchmark more when they can see the peer range, not just a single average, and when their own numbers never leave their machine.
What's next
Add 2021-2024 data, sector-specific metrics (e.g. revenue per employee, deferred-revenue days for SaaS), and a PDF export.
Built by Dr. Muhammad Imran Bhatti, PhD (FCMA, FCPA) - fractional CFO and AI-for-finance builder. Code: https://github.com/drimranbhatti/global-firm-financials
Built With
- github
- html5
- huggingface
- javascript
- matplotlib
- pandas
- python
- scikit-learn
