Free sample datasets for practice
14 ready-to-use datasets for Excel, SQL, Pandas, Power BI and machine learning. Each file is generated fresh when you click, in CSV, Excel, JSON or SQL, with up to 10,000 rows. All the data is fictional and nothing is stored.
Each page explains the columns, shows a preview and suggests exercises.
100 to 10,000 rows, as CSV, Excel, JSON or SQL. Add real-world mess if you want to practise cleaning.
The rows are made up on the spot and sent straight to you. Nothing is saved on our side.
Sweet Shop Sales
Two years of orders from a fictional chain of Indian sweet shops, with Diwali and Raksha Bandhan rushes.
12 columns · Excel pivot tables · Pandas groupby · Time series · Power BI
Employee & Attrition Records
An HR file of employees with salary, satisfaction, overtime and whether they left: built for attrition analysis.
14 columns · Classification · Feature analysis · Excel lookups · Data visualisation
Student Exam Results
Marks in five subjects with attendance and study hours, plus totals, percentage, result and CBSE-style grade.
16 columns · Excel formulas · Correlation · Regression · Conditional formatting
Customer List (CRM)
Customer records with contact details, segment, orders and spend: ideal for cleaning, deduplication and RFM.
13 columns · Data cleaning · Segmentation · VLOOKUP / XLOOKUP · SQL
Stock Register (Inventory)
Monthly stock register for five warehouses: opening, received, issued, closing, reorder level and value.
15 columns · Excel formulas · Data validation · SQL joins · Inventory analysis
Website Traffic (Analytics)
Daily sessions, users, bounce rate and conversions by page, traffic source and device.
10 columns · Dashboards · Time series · Funnel analysis · Power BI
Daily City Weather
Simulated daily temperature, rain, humidity and AQI for Indian cities, with monsoon and winter smog seasons.
9 columns · Time series · Matplotlib · Seasonality · Resampling
Stock Prices (Fictional Tickers)
Daily open, high, low, close and volume for made-up companies (tickers start with SIM): practise returns, moving averages and charts.
9 columns · Financial analysis · Moving averages · Volatility · NumPy
Bank Statement Transactions
Personal bank statements with salary, rent, bills and UPI spends, plus a running balance.
9 columns · Budget analysis · Categorisation · Running totals · Excel
Loan Applications
Loan applications with income, credit score, loan size and term, and whether each was approved.
15 columns · Classification · Feature engineering · Logistic regression · Model evaluation
Movie Ratings
Users rating fictional films from 0.5 to 5 stars: build your first recommendation system.
8 columns · Recommender systems · Pivot / matrix ops · Collaborative filtering · SQL aggregates
Machine Sensor Readings (IoT)
10-minute readings from factory machines that slowly wear out, warn and fail: predictive maintenance practice.
8 columns · Anomaly detection · Time series · Rolling features · Classification
Food Delivery Orders
Orders with distance, weather, traffic, prep and delivery times, and the customer's rating.
13 columns · Regression · Feature analysis · Hourly patterns · Visualisation
Customer Satisfaction Survey
Survey answers on a five-point scale, an NPS score from 0 to 10 and short written comments.
12 columns · Survey analysis · NPS · Encoding categories · Text basics
Why generated data instead of a fixed file?
Most practice datasets online are one fixed file that thousands of people have already analysed. Answers get copied, and you never learn what to do when the data changes. Here every download is new. You choose the size, and the same patterns are always built in, so the techniques you learn carry over while the numbers are yours to work out.
It also means we never have to hand out anyone's real information. Real business data contains customers, employees and money, and publishing it, even "for practice", would be a breach of trust. These files only behave like real data: Diwali sales spikes, monsoon rain, employees who leave when they work too much overtime, machines that vibrate more before they fail.
Which format should I choose?
- CSV: the universal choice. Opens in Excel and Google Sheets, and loads into Pandas, R and Power BI.
- Excel (.xlsx): real dates and numbers, a bold frozen header row, and an "About" sheet describing the file.
- JSON: one object per row, for JavaScript, APIs and NoSQL practice.
- SQL: a
CREATE TABLEstatement plusINSERTs. Run it in MySQL, PostgreSQL or SQLite and start querying.
What does "add real-world mess" do?
Clean data is rare at work. Tick the messy option and the file gets the problems people actually meet: a few blank cells, repeated rows, dates written in several formats, names with stray spaces or odd capitals, amounts saved as text with a ₹ sign, the occasional typing slip that adds a zero, and phone numbers in mixed formats. Cleaning that up is often most of the job, and the Pandas and Excel tracks show you how.
Frequently asked questions
Is this real data?
Do you keep a copy of the files?
Can I get exactly the same file again?
seed4821). Enter the same seed, row count and options and you get the same rows again, which is handy for classes where everyone needs identical data.Can I use the data in my course, blog or project?
Why is the limit 10,000 rows?
Can I load a dataset straight into Python?
pd.read_csv() call with the download link, so you can pull fresh data straight into a Jupyter notebook or Google Colab.