Concept
1. Why projects beat certificates
A recruiter can't see "Advanced Excel" on a certificate. They can see a dashboard that found a problem and recommended an action. Two finished projects beat five half-done ones — pick the 2–3 closest to your target role.
| Target role | Best projects |
|---|---|
| Data Analyst | Retail, E-commerce funnel, HR attrition |
| MIS Executive | Retail, Inventory |
| Finance | Personal finance (+ the Finance dashboard from the Dashboards track) |
| Operations | Inventory, Retail |
| HR analytics | HR attrition |
Project 1 — Retail sales performance
Data: retail_sales.csv — 50,000 order lines, FY2024-25 and FY2025-26, 8 stores, 4 regions, 12 products, 16 salespeople.
Business question: How did FY2025-26 perform vs last year and target, and where should we focus?
Tasks: clean & model (Power Query), pivots/measures for YoY and margin, sales dashboard (Dashboards track Build 1).
Deliverables: dashboard, 5-bullet insight summary.
Check: FY2025-26 revenue ₹29.50 cr, +11.0% YoY, margin 16.5%; top salesperson Amit (Karol Bagh).
Project 2 — HR attrition analysis
Data: hr_attrition.csv — 650 employees (EmpID, Department, Gender, Age, JoinDate, ExitDate, ExitReason, MonthlySalary, PerformanceRating, Overtime, City).
Business question: Who is leaving, why, and what can we change?
Tasks: active vs exited, attrition by department / overtime / age band / rating, exit reasons, tenure at exit.
Trap to handle well: "174 of 650 left = 26.8%" is cumulative since 2019, not annual attrition. Annual attrition for FY2025-26 = exits in the year ÷ average headcount = 25 ÷ ((479 + 476)/2) = 5.2%. Saying this difference in an interview is a strong signal.
Check: exits by department (share of all ever hired): Customer Support 34.5%, Sales 32.2% … HR 17.5%; Overtime "Yes" 37.7% vs "No" 22.2%; top exit reason Better pay (64 of 174).
Project 3 — E-commerce funnel
Data: ecom_funnel.csv — daily Jan–Mar 2026 by Channel (5) and Device (2): Sessions, ProductViews, AddToCart, Checkout, Orders, Revenue (900 rows).
Business question: Where do visitors drop off, and which channel/device should we fix first?
Tasks: funnel chart, step conversions, conversion by channel and device, weekday vs weekend, revenue per session.
Check: 8,77,710 sessions → 16,177 orders = 1.84% conversion; mobile 1.60% vs desktop 2.46%; revenue ₹3.23 cr, AOV ≈ ₹1,994.
Project 4 — Personal finance tracker
Data: personal_finance.csv — 467 transactions, Apr 2025–Mar 2026 (Date, Description, Category, Mode, Type, Amount).
Business question: Where does the money go, and what is the real savings rate?
Tasks: monthly income vs expense, category breakdown, UPI/Card/Cash mix, festival-month spike, budget vs actual with targets you set, a self-updating tracker (dynamic arrays).
Insight to find: income ₹10,20,000; all outflows ₹9,62,548 → cash savings only 5.6%, but outflows include ₹1,44,000 of SIP investment, so the real savings rate is (57,452 + 1,44,000) ÷ 10,20,000 = 19.8%. Classifying investment correctly changes the story.
Check: biggest categories Rent ₹2,64,000, Shopping ₹1,86,022, Groceries ₹1,58,351.
Project 5 — Inventory health
Data: inventory_skus.csv (60 SKUs: Category, UnitCost, OpeningStock, ReorderLevel, LeadTimeDays) + inventory_movements.csv (3,969 In/Out movements, Jan–Mar 2026).
Business question: What should we reorder, what is slow-moving, and is the stock data trustworthy?
Tasks: closing stock (opening + in − out), consumption value, ABC analysis (A = top 80% of consumption value), days of stock, reorder list, slow movers.
Trap: 11 SKUs show negative closing stock — an impossible physical balance that may reflect missing receipts, wrong opening balances, duplicate issues or timing errors. Flag them as a data-quality issue instead of reporting them.
Check: ABC split A 16 · B 18 · C 26 SKUs; 13 SKUs at or below reorder level (11 of them are the negative-stock SKUs).
A standard project structure
- Question (1 line) → 2. Data & cleaning (what you fixed) → 3. Analysis (pivots/measures) → 4. Dashboard → 5. Insights & recommendations (3–5 bullets with numbers) → 6. What I'd do next.
Project conventions: Use the included retail_targets.csv for target comparisons (FY2025-26, Region × Month only). Retail sales is one flat file here; the 24-file Power Query exercise belongs to Analysis & Visualization. HR headcount is measured at close of 31-Mar-2025 and 31-Mar-2026: JoinDate ≤ date and ExitDate blank or > date. Overtime comparisons are cumulative associations, not causes. Inventory ABC sorts consumption value descending, SKU ascending for ties; cumulative share including each SKU ≤80% is A, ≤95% is B, otherwise C. Daily usage is Q1 outward quantity / 90 days; negative stock has no meaningful days-of-stock ratio. Personal-finance savings here means retained cash plus investment contributions, not investment returns.
Common mistakes
Copying a tutorial dashboard without a question. No insights, just charts. Ignoring data problems (negative stock, cumulative vs annual). Five unfinished projects.