The finance projects that impress recruiters are a three-statement model of a listed company, a DCF with sensitivity analysis, a five-year ratio dashboard, a portfolio optimisation in Python, a credit scorecard, a payments funnel analysis and a fintech product teardown. Use public annual reports and free data; explain every line.
Use listed Indian companies: annual reports are free on company websites and the exchanges, and interviewers know the businesses.
- 1Pick a companyannual reports, filings
- 2BuildExcel or Python model
- 3Checkbalance, sanity, sources
- 4Analysevaluation, risk, scenarios
- 5Presentone page, assumptions
The 9 projects
Three-statement financial model of a listed company
The question: Can you rebuild a company's income statement, balance sheet and cash flow and project three years?
- Start with
- Five years of annual reports from a listed Indian company's investor page or the exchange filings.
- Steps
- Enter historicals; reconcile totals to the reports.
- Build drivers: revenue growth, margins, working-capital days, capex.
- Link the three statements so the balance sheet balances.
- Project three years with a base, bull and bear case.
- Measure
- The balance sheet balances in every year; every input cites a page.
- Ship
- The model and a one-page summary.
- Why it counts
- The core finance interview artefact.
Built a linked three-statement model for [company] with three-year projections and scenario cases.
DCF valuation with sensitivity tables
The question: What is the company worth under different growth and discount-rate assumptions?
- Start with
- Your model from #1 plus market data for beta and risk-free rate.
- Steps
- Compute free cash flow to the firm.
- Estimate WACC with sourced inputs.
- Discount cash flows and a terminal value; derive equity value per share.
- Build a two-way sensitivity table (WACC vs terminal growth).
- Measure
- Value per share is within a sensible range of the market price and you can explain the gap.
- Ship
- Valuation tab and a short memo.
- Why it counts
- Shows judgement, not just mechanics.
Valued [company] with a DCF and sensitivity analysis; explained the [x]% gap to market price.
Five-year ratio dashboard for three competitors
The question: Which company in the sector is strongest on profitability, leverage and efficiency, and why?
- Start with
- Annual reports of three companies in one sector.
- Steps
- Compute margins, ROE, ROCE, debt to equity, interest cover and working-capital days.
- Chart five-year trends side by side.
- Write three insights backed by the ratios.
- Measure
- Ratios recalculate from reported numbers; definitions stated.
- Ship
- Dashboard and a one-page comparison.
- Why it counts
- Fast to build and easy to discuss in interviews.
Built a 5-year ratio dashboard comparing 3 companies and identified the sector leader on ROCE and leverage.
Portfolio optimisation in Python
The question: What mix of ten stocks gives the best return for a given risk, and how unstable is that answer?
- Start with
- Daily prices for ten NSE stocks via yfinance (check the data terms).
- Steps
- Compute returns, volatility and correlations.
- Simulate thousands of random portfolios; plot the efficient frontier.
- Find the maximum Sharpe portfolio.
- Test on a holdout period and discuss how the weights change.
- Measure
- In-sample vs out-of-sample Sharpe reported honestly.
- Ship
- Notebook with charts and a limitations section.
- Why it counts
- Combines finance with code, which fintech teams want.
Built a mean-variance portfolio optimiser for 10 NSE stocks and evaluated it out of sample.
Credit scorecard with logistic regression
The question: Which applicants are most likely to default, and where should the cut-off be?
- Start with
- UCI Default of Credit Card Clients (30,000 customers).
- Steps
- Explore default rate by limit, age and payment history.
- Bin variables and fit a logistic regression.
- Convert to a points-based scorecard.
- Choose a cut-off from the cost of a bad loan vs a lost good customer.
- Measure
- Gini or AUC on a holdout set and a cut-off justified by cost.
- Ship
- Scorecard sheet and a decision memo.
- Why it counts
- Exactly what lending teams build.
Built a credit scorecard (AUC [x]) and set an approval cut-off from expected loss.
UPI or payments funnel analysis
The question: Where do payment attempts fail, and what would a 1-point improvement be worth?
- Start with
- Simulate 50,000 transactions with realistic failure reasons, or use a public payments dataset; state your assumptions.
- Steps
- Define the funnel: initiated, authenticated, authorised, settled.
- Compute drop-off by step, bank, device and time of day.
- Estimate the revenue impact of fixing the largest drop.
- Recommend two product changes.
- Measure
- Funnel counts are consistent; impact maths is shown.
- Ship
- Dashboard and a one-page memo.
- Why it counts
- Fintech product analytics in one artefact.
Analysed a 50k-transaction payments funnel; identified a [x]% failure cluster and quantified a ₹[y] recovery opportunity.
Fintech product teardown
The question: How does this app make money, manage risk and comply, and what would you improve?
- Start with
- A UPI app, a lending app or an investment app you use.
- Steps
- Map the onboarding and KYC flow with screenshots.
- Identify revenue lines, fees and partners.
- List the regulations that apply (RBI, SEBI, KYC, data protection).
- Propose three improvements with expected impact.
- Measure
- Every claim about regulation links to an official source.
- Ship
- A 4-page teardown.
- Why it counts
- Shows domain understanding beyond spreadsheets.
Wrote a product and regulatory teardown of [app] with 3 improvement proposals.
Personal finance planner with SIP and EMI calculators
The question: Can a student plan a budget, loans and investments in one sheet?
- Start with
- Your own assumptions.
- Steps
- Build a monthly budget with categories and targets.
- Add EMI and SIP calculators with formulas.
- Chart net worth over time under two savings rates.
- Share a template version.
- Measure
- Formulas match standard EMI and future-value maths.
- Ship
- A public template link.
- Why it counts
- Simple, useful, and shows spreadsheet fluency.
Built a shared personal-finance template with EMI and SIP calculators used by [n] people.
Backtest a simple trading rule, honestly
The question: Does a moving-average crossover beat buy-and-hold after costs, or does it just look like it does?
- Start with
- Daily prices for an index ETF and five stocks.
- Steps
- Implement the rule and buy-and-hold.
- Include transaction costs and slippage.
- Test across periods; report drawdowns.
- Write why most simple rules fail out of sample.
- Measure
- Results with and without costs, across at least three periods.
- Ship
- Notebook and a one-page conclusion.
- Why it counts
- Demonstrates scepticism, which finance interviewers value.
Backtested a trading rule across 3 periods with costs; documented why in-sample gains disappeared.
12 more project ideas
Shorter briefs for when you want a different angle. Apply the same rules: a real source, a baseline, a number and a README.
Mutual fund SIP comparison tool
BeginnerSheets or Python
DataPublic NAV history (AMFI)
Stand-outRolling returns, not point-to-point returns
Bank statement analyser
BeginnerPython, pandas
DataYour own statements (keep private)
Stand-outCategory rules with a monthly dashboard
Sector comparison of five IPOs
IntermediateExcel, prospectuses
DataPublic offer documents
Stand-outValuation at listing vs one year later
Working-capital model for a small business
IntermediateExcel
DataA shop that shares numbers
Stand-outCash-conversion cycle and a credit recommendation
GST and invoice tracker
BeginnerSheets, Apps Script
DataSample invoices
Stand-outAutomated monthly summary
Fraud-rule simulator for UPI
IntermediatePython
DataSynthetic transactions
Stand-outRules tuned to a false-positive budget
Interest-rate sensitivity of a bond portfolio
IntermediateExcel or Python
DataPublic bond data
Stand-outDuration and convexity with a shock table
Startup unit-economics model
BeginnerExcel
DataA fictional D2C brand
Stand-outCAC payback and contribution margin
Crypto or equity volatility study
IntermediatePython
DataPublic price data
Stand-outHonest conclusion about predictability
Annual report summariser with checks
AdvancedLLM API, Python
DataThree annual reports
Stand-outEvery extracted number verified against the PDF
Personal tax planner
BeginnerSheets
DataPublic tax rules
Stand-outOld vs new regime comparison with assumptions stated
Where to get free financial data
- Company investor-relations pages and exchange filings for annual reports.
- RBI and SEBI for regulation and statistics.
- yfinance for price history (check usage terms) and Zerodha Varsity to learn the concepts.
- UCI and Kaggle for credit and transaction datasets.
Frequently asked questions
What is a good finance project for a BBA or MBA student?
A three-statement model and DCF of a listed company, plus a ratio comparison with competitors. Add a fintech teardown for product roles.
Do I need Python for finance projects?
No for modelling and valuation; yes for portfolio optimisation, scorecards and backtests, which strengthen fintech applications.
Where do I get annual reports for Indian companies?
From the company's investor-relations page and the BSE or NSE filings pages, free.
How do I present finance projects to recruiters?
One page per project: question, method, key number, assumptions and limits, with the model linked.
Learn this with mentors
Finance programmes with admissions open: live classes, mentors and project work.
