On this page
The hardest choice in a Class 12 Informatics Practices project comes first: the dataset. Pick one that is too big or too messy, and you spend months fighting your own code.
How many marks are the IP project and viva worth?
In CBSE's Informatics Practices curriculum for 2026-27, Class 12 IP has 70 theory marks and 30 practical marks. Within those 30, project work carries 5 marks and the viva-voce carries 5 marks. The rest goes to Pandas and Matplotlib programs (8), SQL queries (7) and the practical file (5), which must include at least 15 Pandas programs, 4 Matplotlib programs and 15 SQL queries.
You can work alone or in a group of 2 to 3 students, and CBSE says to start at least 6 months before the submission deadline. It suggests visiting local shops or organisations to see how data is generated, stored and managed.
How do you choose a dataset you can handle?
Ask five questions before you write any code:
Can you explain every column in plain words?
Does it fit in one CSV you can scroll through in a spreadsheet?
Does it have a category column and a number column? That gives you GROUP BY and a bar graph.
Does it have dates? That gives you a line plot; a spread of values gives you a histogram, the three plot types the CBSE IP syllabus names.
Can you say exactly where it came from? Data you collected yourself, school data shared with permission, or an open source such as data.gov.in (which that syllabus itself suggests) are all easy to reference.
If a dataset fails two of these, drop it. If your own data needs tidying, our guide to cleaning survey data for school projects walks through the usual fixes.
Here is a worked example. Say the school canteen in-charge lets you record a month of sales: date, item, category, quantity and price, with nothing personal such as who bought what. That one table supports a full project:
Pandas: read the CSV, add a revenue column, filter rows, find the top items.
SQL: a MySQL table, then SUM with GROUP BY category, ORDER BY and MONTHNAME().
Matplotlib: daily revenue as a line plot, a bar graph by category, a histogram of order sizes.
The question: which items sell best on which days?
How much scope is enough for a Class 12 IP project?
A safe rule: one dataset, one question, and a handful of features you can explain line by line. Sort your ideas into three lists (our suggestion, not a CBSE rule):
Core: CSV import and export, a menu-driven Python program, a MySQL table with a primary key, and at least one chart of each plot type.
Nice to have: a second table joined with an equi-join (it is in the Class 12 syllabus), or user input that filters the results.
Leave out: login systems, web scraping, machine learning, or any library you cannot explain.
One way to spend CBSE's six months:
Month 1: choose the dataset and question, get permission.
Month 2: collect and clean data, build the CSV and MySQL table.
Month 3: Pandas analysis and SQL queries.
Month 4: charts and the menu program.
Month 5: write the report, test with odd inputs.
Month 6: rehearse the viva; the rest is buffer.
What should your IP project report include?
Follow your school's format; a report that helps you in the viva also has:
the one question you set out to answer
the data source, permission, and what each column means
the code, with comments in your own words
each chart with one sentence on what it shows
limitations, and references for data and images
What questions should you prepare for the IP viva?
Expect questions on your project and on the syllabus. Start with a short explanation you can say without notes:
"Our project analyses a month of canteen sales, collected with permission. We used Pandas and SQL to total sales by item and day, and plotted the results. The main finding is ... One limitation is ..."
Then practise these:
What is the primary key in your table, and why?
How is a Series different from a DataFrame?
When do you use WHERE, and when HAVING?
Why a histogram here and not a bar graph?
If HAVING still feels shaky, the Fermi Concept Library gives each computing concept its own lesson. If you are not sure where to start revising, Fermi's AI tutor can work out what you already know, build a plan for the session and check your understanding as it goes. Our piece on how to speak confidently in a viva covers the nerves.
Pick your dataset this week, run it through the five questions, and write your one-line question at the top of a blank page.
