SGGSWU Data Analytics Internship Teamwork Guide

November 13, 2026

Web Stories

Teams in a short data analytics internship avoid conflict by agreeing, on day one, who owns each task, how analysis decisions get made and when work is frozen. Disagreements over cleaning rules, metrics or chart choices are normal; they turn into conflict only when no one has decided how to settle them. This guide gives Sri Guru Granth Sahib World University (SGGSWU) students a working agreement, an ownership sheet and scripts for those moments, all fitted to a 14-day plan.

Why analytics teams clash in a two-week internship

Analytics work has more judgement calls than most student assignments. Two people can look at the same dataset and honestly disagree about whether to drop outliers, fill missing values or report a mean versus a median. In a long project there is time to test both options. In 14 days there is not, so unresolved choices quietly eat the schedule.

The usual sources of friction are predictable:

Silent overwrites: one person re-cleans the data and the others’ charts no longer match.
Unclear ownership: two people build the same chart while nobody writes the summary.
Undocumented choices: a filter is applied and no one remembers why.
Uneven effort: one member waits for others, or does everything and resents it.
Late disagreement: a basic method dispute surfaces on day 12 instead of day 3.

The fix is structure, agreed early and written down. The rest of this article provides it.

A 14-day team timeline built around decision points

The free two-week online internship from CareerSwami runs for 14 days, so a team plan has to be tight. The table below places the decisions that most often cause arguments on specific days, so they are settled before they hurt.

Days	Team focus	Decision to lock
1	Working agreement and ownership sheet	Roles, channels, meeting times, decision rule
2-3	Understand the data and the question	The one-sentence business question and success metric
4-5	Cleaning and preparation	Missing-value, duplicate and outlier rules (written in a log)
6-8	Exploration and analysis	Which methods and metrics are used in the final story
9-10	Visuals and dashboards	Chart types, colour and labelling conventions
11-12	Findings and recommendations	Final claims; what the data does not support
13	Cross-review	Every number checked by someone who did not produce it
14	Freeze and submit	No new analysis after the freeze time

Note that the cleaning rules are locked by day 5. Most late conflicts come from reopening those rules after charts exist.

Free 2-Week CareerSwami Internship

Online, 14 days, with a free offer letter and a free completion certificate once you finish the work. Check with your college that it fits your requirement before you enrol.

Apply for the free internship

Team working agreement template for a data analytics internship

Copy this into a shared document and fill every blank in a single 30-minute call on day 1. Each member types their name under “Agreed” so nobody can later say they were not consulted.

TEAM WORKING AGREEMENT
Team name: ______  Track/project: ______
Members: ______ (name, city/time zone, best contact)

1. GOAL
Our one-sentence question: ______
Our deliverable: ______ (notebook / dashboard / report)

2. COMMUNICATION
Main chat group: ______
Reply within: ____ hours (agreed quiet hours: ____)
Daily check-in: ____ (time) for ____ minutes, or written update by ____

3. FILES AND VERSIONS
Single shared folder: ______
Raw data is read-only; nobody edits the original file.
File naming: YYYYMMDD_task_initials_v#
Only the data-prep owner changes the cleaned dataset.

4. DECISIONS
Small choices (chart colour, axis labels): task owner decides.
Method choices (cleaning rule, metric, model): owner proposes in
the decision log; team has 24 hours to object.
If still disputed: run a quick test (see script 3), then vote.
Tie-break person: ______ (rotates on day 8)

5. QUALITY
Every number is reviewed by someone who did not produce it.
Assumptions go in the decision log the same day.

6. WORKLOAD
Planned effort: about ____ hours per person per day.
If someone is blocked or unwell, they tell the team within ____ hours
and the owner sheet is updated.

7. FREEZE
No new analysis after: Day 14, ____ (time).

Agreed: ______ ______ ______ ______ (date)

Why the decision log matters

A decision log is a simple table with four columns: date, decision, reason, who agreed. It turns “I thought we were dropping those rows” into a checkable record. It also gives each member material to describe honestly in their own words when they later explain the project.

Task-ownership sheet for a four-person analytics team

Each task gets one owner (accountable) and one reviewer (checks). A person can own several tasks, but nobody reviews their own work. The rows below are an illustrative layout, not a required format; adjust to your team size and track.

Task	Owner	Reviewer	Due day	Status
Problem statement and success metric	Member A	Member B	3	To do
Data dictionary (what each column means)	Member B	Member C	3	To do
Cleaning and decision log	Member C	Member A	5	To do
Exploratory analysis	Member D	Member B	8	To do
Core analysis / metrics	Member A	Member D	8	To do
Charts and dashboard	Member B	Member C	10	To do
Findings and recommendations	Member D	Member A	12	To do
Number check against raw data	Member C	Member B	13	To do
Final packaging and submission	Member A	All	14	To do

Use only four status words: To do, In progress, In review, Done. Update the sheet at each daily check-in so that anyone can see where the team stands in ten seconds.

Scripts for disagreements over analysis choices

These are word-for-word openers. The aim is to move the argument from people to evidence.

Script 1: two people want different ways to handle missing values

“I see two options: drop the rows, or fill with the median. Let’s write both in the decision log with one line on what each could distort. Then we run the main metric both ways and see whether the answer changes. If it doesn’t, we take the simpler one.”

Script 2: a teammate changed the cleaned data without telling anyone

“My charts stopped matching the table, and I think the cleaned file changed. Can we agree that only the data-prep owner edits it and announces changes in the chat? For now, could you tell me exactly what changed so I can update my part?”

Script 3: you disagree about whether an outlier is an error or a real value

“Let’s not argue about it in the abstract. Can we check whether the value is possible in the real context, and see what the result looks like with and without it? Whichever we choose, we state it in the report.”

Script 4: a teammate insists on a complex method you think is unnecessary

“What would this method let us say that the simpler summary can’t? We have 14 days, so I’d like to try the simple version first and use the complex one only if it changes a recommendation.”

Script 5: someone is not delivering

“Your part on the sheet is due on day 8 and it’s still marked To do. Is something blocking you? If it helps, I can take the first half and you take the review, so we stay on schedule.”

Script 6: the team is deadlocked

“We’ve each made our case. Per our agreement, we vote now and the tie-break person decides if it’s split. Whatever we choose goes in the decision log, and we move on without reopening it.”

Worked example: settling a dispute over a sales-by-region chart

This example is illustrative. The numbers and names are invented and no code was run.

A four-person team analyses a retail sales dataset. On day 7, Meera wants to show average order value by region. Harpreet argues that a few very large orders inflate the average in one region and wants the median.

Name the disagreement: “We disagree on mean versus median for order value.” This is a method choice, so it goes to the decision log.
State the question the chart answers: “Which region has the highest typical order?” The word “typical” favours the median, but the team notes that total revenue questions would favour the sum.
Test, do not debate: each person computes both values for every region. If the region ranking is the same, the dispute is settled by simplicity. If it differs, the team shows both in the report with a note on why.
Record it: the log entry reads “Day 7: used median order value for typical-order comparison; mean shown in appendix; reason: a handful of large orders skew the mean; agreed by all four.”
Move on: the dashboard owner updates the chart label to say “Median order value” so readers are not misled.

The whole process takes a short call, not a week of tension, because the agreement already said what to do.

Checklist for the daily team check-in

Keep the call to 15 minutes. Each person answers three questions in order, and nothing else is discussed until all have spoken.

What did I finish since yesterday, and where is the file?
What will I finish today?
What is blocking me, and who can unblock it?

Then review the ownership sheet, check the decision log for new entries and note any disagreement that needs the 24-hour objection window.

SGGSWU students: fitting a team internship into university rules

SGGSWU’s Self Study Report describes the university as a State Private, Unitary university, with no constituent, affiliated or autonomous colleges. In practice, that means rules on internships and training come from the university itself, not from a separate affiliating body. The same report says mandatory practical-based internships and projects are part of its “Learning by Doing” approach and that 46 of 56 programmes have field project, research project or internship components. For students in computing programmes such as BCA, MCA or BTech CSE, which the RTI declaration lists, analytics projects are a natural fit.

The University Calendar 2024 says students must take all papers, including Project/Summer Training, as per the approved scheme for their course, under a credit and grade point system. A voluntary online internship is not automatically the same as that required component, and no programme-specific rule on online internships was confirmed. Do not assume it counts toward anything; ask first. The 14-day internship from CareerSwami offers a free offer letter and a free completion certificate after successful completion, but no stipend, academic credit or placement.

The Training and Placement Cell page (excerpt-based, so check the current version) describes the cell as facilitating internships, industrial training and live projects. Your team can use this as a point of contact. Verification steps:

Find your programme’s approved scheme and check what the Project/Summer Training component requires.
Email your department coordinator asking whether an external online internship can be recorded or recognised, and request the answer in writing.
Ask the Training and Placement Cell what documents, if any, they want for external internships.
Check with your faculty whether group work on one dataset is acceptable for anything you plan to submit to the university.
Keep your offer letter, certificate and decision log in one folder in case the department asks for proof of work.

Warning signs that a team is drifting, and the fix for each

Sign	Likely cause	Quick fix
Chat is busy but the sheet has not changed	Talking instead of deciding	Move the topic to the decision log with a 24-hour deadline
Two versions of the same chart	Unclear owner	Name one owner; the other becomes reviewer
One person goes quiet	Blocked, unwell or overloaded	Private, friendly message using Script 5
Numbers differ between report and dashboard	Edited cleaned data	Rebuild from raw data; restrict edit rights
Debates restart on day 11	Decisions were never logged	Point to the log; reopen only with new evidence

Giving credit fairly in a shared analytics project

Arguments about credit are easier to prevent than to settle. Use the ownership sheet as the record of who did what, and describe your own contribution accurately when you talk about the project afterwards. Say “I owned the cleaning and the decision log” rather than “we built everything”. Do not claim results the data did not support, and do not describe teammates’ work as yours.

Frequently Asked Questions

How long should the team working agreement be?

One page is enough. Cover the goal, communication, file rules, how decisions get made, quality checks and the freeze time. A longer document tends not to be read.

What if two teammates are equally senior and cannot agree on a method?

Run a quick test where possible: compute the result both ways and see if the conclusion changes. If it does not, pick the simpler option. If it does, report both with a note. If still stuck, follow the vote and tie-break rule you wrote on day 1.

Should everyone work on the same notebook?

Usually not at the same time. Keep raw data read-only, give each person their own notebook or section, and let one person merge into the shared version. This avoids silent overwrites.

How do we handle a teammate who is not contributing?

Start with a private, specific message pointing to the task and due day on the sheet, and offer a way to rebalance work. If nothing changes, raise it in the team call using the sheet as the neutral record rather than making it personal.

When should the team stop analysing and start writing?

Around day 11 in a 14-day plan. New analysis after that point risks mismatched numbers. Use days 13 and 14 for cross-checking and submission only.

Do I need my university’s permission to do an external online internship as a team?

SGGSWU rules on this were not confirmed for any programme. Write to your department coordinator and ask whether external online internships are recognised, and keep the reply. Do not assume it counts toward your required Project/Summer Training.

How do I explain my role in a team analytics project to an interviewer?

Use your ownership sheet and decision log. Name the tasks you owned, one decision you argued for with evidence, and one you changed your mind on. Be specific and honest about what the team did together.

Is the CareerSwami internship suitable for practising this kind of teamwork?

The free 14-day online internship is short enough that a team agreement and decision log make a visible difference, because there is no slack to absorb unresolved disputes. It offers a free offer letter and a completion certificate after successful completion.