Datasets built from real business problems, not random CSVs.
KOJOH is a dataset marketplace for people building a data career. Every dataset starts as an actual problem pulled from Logistics, Sales, Service, Product, E-commerce, or HR, with its own schema, its own story, and a challenge worth solving.
Datasets available now
One-time payment per dataset. Your first paid dataset is at the offer price; later ones at MRP. Instant download with documentation and a project challenge.
Kirana Store Sales
A year of bills from a Lucknow general store. Find what makes money and how much is stuck in credit.
FreeView dataset → Product AnalystThe Funnel That Leaked
An app onboarding funnel and an A/B test where the dashboard gets the winner wrong.
FreeView dataset → Data AnalystWhere Did My Salary Go
Twelve months of UPI and bank statements from 24 first-jobbers. Find subscriptions and double charges.
The Diwali Discount Trap
14,700 e-commerce order lines across four marketplaces. Why did the biggest sales month lose money?
Late Again: Last-Mile Delivery
A regional courier's dispatch data. Find where and why deliveries miss their promise date.
Spares Never Arrive
Appliance service tickets and spare-part orders. Track turnaround time and the parts that cause delays.
The Quiet Quitters
Streaming app subscriptions and viewing logs. Predict who churns before they cancel.
Stockout Saturdays
92,000 store-by-SKU days from a quick commerce app. Forecast demand when empty shelves hide it.
The ₹176 Lakh Question
Hand-typed ERP entries from an 8-branch car dealer group. Who really performed, and what did December cost?
Slow Boat from Ningbo
An import procurement tracker for products sourced from China. Track shipments, pending documents and payments.
Approved Only: Credit Scoring
Personal loan applications and repayments at an NBFC. Build a credit model without falling for leakage.
What KOJOH does
Real, independent datasets
Every dataset keeps its own schema instead of being forced into one generic table, because real business data never looks the same twice.
Organized by job track
Browse by the role you're building toward, not just a keyword search. Each folder is a career path, not a category label.
A challenge, not just a file
Each dataset comes with the business context behind it (what question it's meant to answer), so your portfolio shows judgment, not just SQL.
Search that understands the job
Type a role, a tool, or a domain and get matches, powered by an ontology of job titles, skills, and synonyms, not a black-box AI.
Built for four job tracks
Pick the track you're aiming for at signup. It's how your recommendations get tailored.
Data Analyst
Sales, e-commerce and personal-finance data to clean, analyse and turn into a recommendation.
Data Scientist
Churn prediction, demand forecasting and credit scoring, with the traps real models fall into.
Business / Ops
Logistics, service and procurement data where the question is what to fix first.
Product Analyst
Funnels and A/B tests with results that are less obvious than the dashboard says.
New & upcoming
Accounts & job-track profiles
Sign up, pick the professions you're targeting, and get a profile that carries into your recommendations.
Ontology-powered search
Search by role, tool, or domain and get ranked, relevant datasets. No AI black box, just a transparent semantic layer.
More datasets, every week
New job-track datasets are added regularly as the catalog grows beyond the first release.