Models, analysis and writing: shown, not told.
Freelance financial modeling, data analysis and finance writing. Every engagement starts with a short brief and ends with working files you own, not screenshots. The case studies below are live: open one and drag a slider.
- Fixed price per project
- Quote within one business day
- You own the files
- NDA on request
Every piece here is interactive, running its own arithmetic in your browser. Nothing is a picture of a spreadsheet: change an input and every number downstream of it moves, which is the only honest way to show that a model works.
Three-statement model: DSE-listed manufacturer
Linked income statement, balance sheet and cash flow, with a scenario switch and a balance check that is computed rather than asserted: if the model were wrong, it would say so. Drag any assumption and all three statements, the charts and the credit metrics move with it.
- FY24A to FY29E
- Three scenarios
- Live balance check
DCF with sensitivity tables
WACC built up from its parts, two terminal value methods that quote each other back, and a two-way grid on WACC and terminal growth where every cell is a full revaluation and clicking one adopts it.
- WACC built up, not typed in
- Two terminal methods
- Fed by the model above
Index returns: volatility and drawdowns
Rolling volatility, drawdown episodes with their recovery times, fat-tail diagnostics against a normal curve, and what actually happened to someone who held for a day, a year or five.
- Six linked views
- Your CSV, parsed in the browser
- Demo series marked as simulated
A screened FTSE 250 fund, built in 2015 and held to 2020
Ten mid-cap holdings past a Shariah and sustainability screen that runs before any price is looked at, weighted at the end of 2015 by minimising contributed variance, then held unchanged through five years nobody had seen. The frontier and the optimised alternatives are solved beside it, live.
- Ten holdings, £10m
- Weighted 2015, held to 2020
- Computed from daily closes
Probability of default: scorecard vs gradient boosting
The whole pipeline on a real public dataset, fitted in your browser while you read: split, encode, a logistic scorecard against a boosted ensemble, cross-validation, calibration, and the cut-off where the modelling stops and the lending decision starts. Including an honest answer about which model wins.
- Nothing precomputed
- Refits on every seed change
- Calibrated, then priced
Portfolio stress testing: macro shocks to capital
Unemployment, growth and rates driving default rates through a Merton model and a vintage hazard model at once, then loss given default as an option on the collateral, IFRS 9 staging, and the capital ratio. The gap between the two engines stays on the page, because that gap is the model risk.
- Two engines, side by side
- IFRS 9 staging to CET1
- Reverse stress test
Islamic vs conventional funds: an MSc dissertation
The hypothesis was that Shariah-compliant funds carry less risk. It failed, and the power analysis then showed why a sample of three could never have tested it. Every table and series on the page is the submitted one.
- 220 UK equity funds
- 19,577 fund-months
- Five-factor and Carhart
Also on this site
Work that lives somewhere other than a case-study page.
Three services, one standard across them: the working file is the deliverable, the assumptions are visible, and nothing arrives locked.
A model you can edit
Three-statement models, DCF and comparable-company valuations, budgets and scenario models, built in Excel with a clean assumptions tab so you can change the inputs yourself.
What you receive- One assumptions tab, colour-coded: blue inputs, black formulas, nothing hard-coded
- Three linked statements that balance, with the check row visible
- Scenario switch: base, upside, downside, without rebuilding anything
- Sensitivity tables on the two inputs that matter most
- A one-page summary a non-finance reader can follow
- A walkthrough call, so the model outlives the engagement
An answer, plus the working
Cleaning, analyzing and visualizing financial or business data in Python, R or Excel. Dashboards and summary decks that answer a specific question rather than displaying everything at once.
What you receive- The question stated precisely before any code is written
- Cleaning steps documented: what was dropped, and why
- Commented Python or R you can re-run when the data updates
- Charts that survive being printed in black and white
- Findings written in sentences, with the caveats attached
- Raw outputs handed over, not just the pretty version
Copy that survives an expert reader
Explainers, market write-ups, research summaries and educational content. Accurate, sourced, and written in English or in everyday Bangla rather than textbook Bangla.
What you receive- English or Bangla, in your house style
- Every factual claim sourced, with links in the draft
- Jargon either explained on first use or removed
- Structured for skim-reading: headings that carry meaning
- Two revision rounds included as standard
- Compliance-safe phrasing where a regulator might read it
The brief
Send what you need with any files or examples, through the contact form or straight to i@reiad.co.uk. A paragraph and a spreadsheet is enough to start.
The quote
A reply within one business day with scope, a delivery date and a fixed price. The price does not move unless the scope does, so the estimate is my problem rather than yours.
Delivery
The working files (model, code or draft), a walkthrough of how they fit together, and one round of revisions included. Ownership transfers on final payment.
Prefer the protection of a platform? I also take projects through Fiverr and Upwork. Get in touch and I will share the current profile links.
One person, start to finish, and I answer my own email.
I am Rony Reiad. Economics at the University of Chittagong, then an MSc in Finance and Risk Management at the University of Brighton, finishing with an Upper Merit, and CFA Level 1 in the November 2026 window. No agency, no account manager, no handoff to someone you have not spoken to.
Everything on this site is my own build: the case studies above, the calculators in the Tools hub, and a Bangla learning library that exists because financial education in Bangladesh usually fails people at the language step rather than the maths step. It is the closest thing to a work sample a first email can offer.
- Excel
- Three-statement
- DCF
- Credit analysis
- Python
- R
- Stata
- SPSS
- EViews
- Bloomberg Terminal
- DSE filings
- Time-series
How much does it cost?
Fixed price per project, quoted after the brief, not hourly, so the estimate is my problem rather than yours. Scope, delivery date and price come in one reply, and the price doesn't move unless the scope does. Send the brief and you'll have a number within a business day.
How long does it take?
It depends on the state of the data far more than on the size of the model. A clean three-statement build is days; the same job where the numbers arrive as scanned PDFs is weeks. The quote states a date, and if something threatens it you hear about it early rather than late.
Who owns the work?
You do: files, formulas, code, all of it, on final payment. No locked spreadsheets, no "contact me to unlock the assumptions tab". If I want to reference the work publicly I'll ask first, and anonymise if you'd rather.
Will you sign an NDA?
Yes, routinely, and I'd rather sign one than have you send a redacted brief that makes the work worse.
What won't you do?
I won't produce a valuation designed to reach a number decided in advance, write promotional copy dressed as independent research, or recommend specific securities to retail readers. Turning down that kind of work is the reason the rest of it is worth anything.
Can I see the code or the model first?
You already can. The case studies above are the samples: open one, drag the assumptions, download the CSV, read the notes on where each number came from. They run the same arithmetic a client build would, on companies and data small enough to publish. For work that sits under an NDA, ask and I'll walk you through a redacted example on a call.
Send the brief: however rough it is
A paragraph and a spreadsheet is plenty to start. You'll get scope, a delivery date and a fixed price back within one business day, with no obligation and no sales sequence.