Preparing for a Data Science Internship as an AKTU Student: Python, Statistics and SQL Checklist

You are ready for an entry-level data science internship if you can load a messy CSV in Python, summarise it with basic statistics, and answer business questions with SQL joins and aggregations, all without copying code line by line. The grid below turns that into a score you can check in an hour and a practice task for every gap it shows.

This guide is written for students of Dr. A.P.J. Abdul Kalam Technical University (AKTU) and its affiliated colleges. It is an independent guide and has no connection with the university.

What a beginner data science internship actually asks of you

Short internships, especially two-week ones, rarely expect advanced machine learning. They expect you to be fluent in the unglamorous layer under it: getting data in, cleaning it, describing it, and explaining what it says. Most internship tasks at this level fall into four activities.

• Read and clean data: fix types, handle missing values, remove duplicates, standardise text.
• Describe data: means, medians, spreads, group comparisons, simple charts.
• Query data: pull the right rows with SQL, join tables, aggregate.
• Communicate: write three or four sentences that a non-technical reader can act on.

Python, statistics and SQL map directly onto these. Machine learning models are a fifth step, and it sits on top of the first four. A student who can do the first four well and train a simple model will usually look stronger than one who can import a classifier but cannot explain why a column has 30% blanks.

How to use the self-assessment grid

Score each row from 0 to 3 using the definitions below. Be strict: a 3 means you did it recently, from a blank file, without looking up the syntax more than once or twice.

• 0: I have not seen this or cannot explain it.
• 1: I recognise it, but I need a tutorial open to do it.
• 2: I can do it with occasional documentation lookups.
• 3: I can do it from scratch and explain my choices aloud.

Add the scores per section. Each section has 8 rows, so the maximum is 24 per section and 72 overall. The interpretation bands come after the grid.

The prerequisite grid: Python, statistics and SQL

Python (8 skills)

Statistics (8 skills)

SQL (8 skills)

Reading your total

Treat the weakest section as your priority, even if you enjoy another one more. Internship tasks tend to expose the weakest link, because one analysis often needs a SQL query, a pandas clean-up and a statistical statement in sequence.

A worked illustrative example: scoring and fixing one student profile

The following is an illustrative scenario, not a real student. Code is shown as a teaching sketch and has not been run here, so test it on your own data.

Imagine a third-year B.Tech student, call her Student A, with these totals: Python 17, Statistics 9, SQL 12. Her overall score is 38 out of 72. The bands say Python is ready, SQL is workable, and statistics is the gap. Her plan follows from that, with about half her time going to statistics.

She picks a small public retail-style dataset with a customers table and an orders table. Her first task is to answer: Do customers in metro cities spend more per order than others? This uses all three skills.

Step 1, SQL: pull one row per order with the customer’s city type.

SELECT o.order_id,
       o.order_value,
       c.city_type
FROM orders o
LEFT JOIN customers c
  ON o.customer_id = c.customer_id
WHERE o.order_value IS NOT NULL;

Step 2, Python: load the result, check for missing city types and compare groups.

import pandas as pd
from scipy import stats

df = pd.read_sql(query, conn)
print(df["city_type"].isna().sum())
df = df.dropna(subset=["city_type"])

summary = df.groupby("city_type")["order_value"].agg(["count", "mean", "median", "std"])
print(summary)

metro = df.loc[df.city_type == "metro", "order_value"]
other = df.loc[df.city_type != "metro", "order_value"]
print(stats.ttest_ind(metro, other, equal_var=False))

Step 3, statistics: this is where her low score shows. She must decide whether the mean or the median is the better summary, since order values are usually right-skewed. She must also read the p-value correctly. A small p-value would suggest the difference is unlikely under the null hypothesis. It would not show that living in a metro causes higher spending.

Step 4, communicate: she writes a short summary in this shape: the data covers N orders; the median order is higher or lower in metro cities by a stated amount; the test suggests the gap is or is not distinguishable from noise; and one limitation, such as customers with a missing city being dropped. That four-part pattern is reusable in almost any internship report.

Student A then re-scores. If S4, S6 and S7 each move from 1 to 2 after a week, her statistics total rises by 3 and the grid has done its job: it showed her where to spend time.

A 14-day preparation roadmap

If you have two weeks before an internship starts, this schedule fits around lectures. Each day is about 90 minutes.

1. Day 1: Take the full grid and record scores. Pick one small dataset you will use all fortnight.
2. Days 2-3: Python and pandas, rows P5 to P7, on your dataset.
3. Day 4: Plotting (P8). Produce three charts with a one-sentence takeaway each.
4. Days 5-6: Descriptive statistics and outliers (S1 to S3), computed by hand once and then with code.
5. Day 7: Rest, and re-do any task you got stuck on without notes.
6. Days 8-9: SQL fundamentals and joins (Q1 to Q5). Load your dataset into SQLite or another local database.
7. Day 10: CTEs and window functions (Q6, Q7).
8. Days 11-12: Inference and regression (S6 to S8) on your dataset.
9. Day 13: Mini-project: one question answered with SQL extraction, pandas cleaning, a statistical check and a chart.
10. Day 14: Write a one-page summary and re-score the grid.

Common mistakes that cost beginners marks and credibility

• Dropping missing values silently. Always report how many rows you removed and why.
• Using the mean on skewed data. Check the median alongside it.
• Joining without checking row counts. A join that multiplies rows quietly inflates every total. Compare counts before and after.
• Confusing WHERE and HAVING. Filter rows first, then filter groups.
• Treating p-values as proof. Use careful wording such as “suggests” or “is consistent with”.
• Training a model before understanding the data. Profile the columns first.
• Uncommented notebooks. A reader should follow your reasoning without you in the room.

AKTU students and the Python, statistics and SQL checklist: what the scheme and the college need from you

AKTU is an affiliating university set up in 2000 by the Government of Uttar Pradesh, with jurisdiction over the whole state. Its admission portal says it affiliates 700+ colleges (a later paragraph gives 756). Two students in the same branch can therefore sit in different colleges with different local timetables and internship coordinators. The grid in this guide is a skills check, not a list of AKTU topics, so tick it against your own branch syllabus.

Timing matters most for a data science internship. An AKTU B.Tech 2nd-year scheme for Plastic Engineering (session 2023-24) states that a 4-week Mini Project or internship is done in the summer break after semester 4 and assessed in semester 5. The Textile Technology scheme lists this as paper BCC351, worth 100 marks and 2 credits. A third-party mirror of the CSE-family scheme (covering Computer Science, IT and AI branches) repeats the same rule. It is not an official page, so open the official copy for your branch. An older 2019-20 scheme, seen only in a third-party copy, placed a 3-4 week project after semester 2. The rule has changed between schemes, so do not rely on a senior’s account.

In practice, a 14-day roadmap on Python, statistics and SQL fits best inside that summer window or a gap in your timetable. Whether a free online internship can satisfy the credit-bearing requirement is not settled by anything you have read here. Treat it as a question for your college.

For guidance, the state’s URISE portal has a Central Training & Placement Cell page for AKTU. In 2025 it carried industry announcements and a virtual internship offer for B.Tech, M.Tech, MBA and MCA students. Check the page for what is current, and approach your own college’s training and placement cell first.

Checks before you count this internship toward AKTU credit

• Download the official scheme PDF for your branch and admission year from the AKTU site, and find the internship or mini-project paper, its marks and its semester.
• Email your internship coordinator or head of department asking whether an online data science internship of this length can be used for that paper, and in what format.
• Ask what documents are needed for assessment, such as a certificate, report, or faculty mentor sign-off, and the deadline in semester 5.
• Match the grid to your DBMS, programming, and probability and statistics papers, and note which rows are self-study.
• Ask your college for its academic calendar for 2026-27 and plan the roadmap around exams and practicals.
• Keep every written reply with its date or circular number.

Turning your preparation into proof

Internships reward visible work. Build a small portfolio as you go.

• One GitHub repository with a README stating the question, the data source, the method and the finding.
• A SQL file with commented queries for the same question.
• A notebook that runs top to bottom without errors from a fresh kernel.
• A one-page write-up in plain language with one chart.

If you want a structured, no-cost way to practise, CareerSwami offers a free two-week online internship across 26 tracks, with a free offer letter and a free completion certificate after successful completion. It does not come with a stipend or academic credit, so treat it as practice and a portfolio-building opportunity.

Frequently Asked Questions

How much Python do I need before a beginner data science internship?

Aim for a score of at least 2 on every Python row, particularly P5 to P7. You should be able to load a CSV, clean it, group it and plot it without a tutorial. Advanced topics such as decorators or multithreading are rarely needed.

Do I need to learn machine learning first?

No. Cleaning, description, SQL and basic inference are the foundation. If you finish the grid early, add linear regression, logistic regression and a train-test split, and learn to explain their results plainly.

Which statistics topics matter most for internship tasks?

Descriptive statistics, outlier handling, sampling, confidence intervals, a two-group comparison and simple regression come up most. Practise explaining each in one sentence to someone outside your branch.

Which SQL database should an AKTU student practise on for a data science internship, given DBMS lab coursework?

SQLite is the easiest because it needs no server setup. MySQL and PostgreSQL are also good choices. Core syntax is mostly shared, though date functions and a few details differ, so check the documentation of whichever one you use.

How long does it take to move from a score of 36 to 55?

That depends on your starting point and hours available. For many students, the 14-day roadmap above, at about 90 minutes a day, is enough to move a few rows from 1 to 2. Re-score on day 14 and extend the plan if needed.

Can I prepare using only my college coursework?

Coursework may cover part of the grid, but practical skills such as pandas cleaning and window functions often need extra hands-on practice. Compare your syllabus with the grid and fill the gaps with exercises on real datasets.

What datasets are good for practice?

Choose one with at least two related tables, some missing values and a clear business question, such as retail orders, ride bookings or student results. Public repositories and open government data portals are good places to look. Check each dataset’s licence before reusing it in a portfolio.

Will an online internship certificate help my AKTU record?

Whether it counts for anything academically depends on your college’s rules, so ask your internship coordinator using the written question in the checklist above. A certificate alone does not guarantee credit, and it works best alongside a project you can explain.

Related Data Science guides

• How MAKAUT Students Can Evaluate a Data Science Internship Offer Before Accepting It
• Building a Data Science Internship Portfolio for GTU Students: Notebooks, Datasets and Write-ups
• Data Science Internship Report Structure for Savitribai Phule Pune University Students
• All Data Science internship guides