Data career plan from second year: a semester-by-semester schedule

The budget before the plan

Most study plans fail because they assume time you do not have. Before copying anything, fill in your own numbers:

The schedule assumes roughly 14 teaching weeks per semester, with the last three at reduced self-study because of exams. That gives about 90 to 110 self-study hours per semester, plus vacations. If your honest number is lower, stretch each block rather than skipping one.

Semester 3: Python you can actually write

Aim to finish the semester able to read a CSV, clean it and summarise it without copying from a tutorial.

• Weeks 1–5: Python basics — variables, lists, dictionaries, loops, functions, files.
• Weeks 6–10: pandas — filtering, groupby, merges, handling missing values.
• Weeks 11–14: one small analysis of a dataset you choose (cricket scores, your hostel mess menu, public transport timetables). Put it on GitHub with a README.

Checkpoint: you can explain every line of your analysis notebook.

Semester 4: SQL and statistics together

These pair well because SQL gives you the numbers and statistics tells you whether to believe them.

• SQL: SELECT, WHERE, GROUP BY, JOINs, then window functions. Practise on SQLite, which needs no installation beyond Python.
• Statistics: mean, median, spread, distributions, sampling, confidence intervals, the idea of a hypothesis test. Your college course may cover this; if it is theory-only, reproduce each concept in Python.
• Summer vacation after semester 4: your first internship attempt, even a short one. See internships during college for what a first internship should prove.

Checkpoint: you can write a query that joins two tables and defend why the row count is right.

Semester 5: one serious project

This is the project you will talk about in interviews for the next two years, so choose a question, not a dataset. "Which bus routes in my city run late most often, and at what time of day?" is a question. "Analysis of a transport dataset" is not.

• Weeks 1–3: define the question, find or build the data, write down what a useful answer looks like.
• Weeks 4–10: clean, analyse, build one dashboard or report.
• Weeks 11–14: write it up, including what went wrong and what you would do next.

Also start a simple log of every job description you read that interests you, with the tools it asks for. By semester 7 this log tells you what to prioritise.

Semester 6: machine learning with evaluation

Learn regression, classification, train/test splits, cross-validation and metrics such as precision and recall. The skill interviewers probe is evaluation, not model names. Build a second project where you compare a simple baseline with one model and explain why the difference matters for a decision.

Summer vacation after semester 6: the internship that counts most. Many companies use summer internships in the pre-final year to decide on pre-placement offers, so apply early in semester 6, not after exams.

Semester 7: placement preparation

Placement season timing varies by college, so check your training and placement cell's calendar at the start of semester 7 and shift this table to match.

Semester 8: a final-year project that survives questions

Your final-year project should be defensible, not flashy. Use it to show depth in one area you want to work in. The rubric in final-year data project ideas helps you check this before your viva.

What to do if you are starting in semester 5

Compress semesters 3 and 4 into one: Python and SQL in parallel, statistics only as far as confidence intervals. Skip the second project and make the first one strong. You lose breadth, not readiness for entry-level analyst roles.

A tracking sheet you can build in ten minutes

Make one row per week with these columns: week number, planned hours, actual hours, topic, one thing you can now do, one thing still confusing. Review it at the end of each month. The "still confusing" column becomes your revision list before placements, and the "actual hours" column tells you honestly whether the plan fits your life.

If you later want mentor feedback on projects and a structured path from Python to deployed models, the Data Science programme covers that sequence with live, recorded sessions.