Phase 01 · Foundations · pandas
Data Analysis Foundations
Turning a vague business question into a checked, honest number.
Learn to think like an analyst: turn a vague question into a table question, check the data, compare groups fairly and write a finding. All in the pandas editor, no setup.
Why this track exists
Someone sends you one line: "Is UPI a problem?" You could open a table and start counting, but a count alone proves nothing. Is 12 failures a lot? Compared with what? Out of how many? Over which dates? Do the two tables that hold this money even agree? Analysts get paid for the part before and after the code: asking the right question, checking the data, comparing groups fairly, and saying how sure they are. This track teaches that habit, using pandas as the tool.
In a real analyst or data job, most of the day is not writing clever code. It is clarifying a request, checking that numbers reconcile, picking between a mean and a median, and writing three honest lines for a manager. The code is short. The judgement is the skill.
What you need before starting
Every lesson runs in the editor with the data already loaded. If you have never used Python, the first two lessons still work, and the Python track fills gaps.
The Pandas track teaches the tools in more depth. This track shows how an analyst uses them.
The roadmap
3 stages, in the order they build on each other. Each stage lists the modules and lessons it covers, and what you should be able to do by the end of it.
01Think like an analyst
Most wasted analysis comes from a question nobody pinned down. A vague brief produces a confident answer to the wrong question.
Think like an analystFree
0/2
Turn a vague question into a table question, then take a first look at a dataset before you trust it.
By the end of this stage
You can turn a one-line request into a metric, rows, a comparison and a time range, and take a first look at an unfamiliar table before trusting it.
Where people get stuck
Skipping the first look. Two minutes of shape, nulls and value_counts catch problems that take hours to find later.
02Trust the numbers
A number is only as good as the checks behind it. Two sources that should agree often differ, and one big order can move an average.
Trust the numbers
0/3
Check that two sources agree, pick the right kind of average, and decide what to do with extreme values.
By the end of this stage
You can reconcile two tables, choose between mean and median, and measure how much extreme rows matter before you decide what to do with them.
Where people get stuck
Deleting outliers because they look odd. Measure and investigate first, then decide and write the reason down.
03Compare, spot trends, report
Decisions come from comparisons, and comparisons are easy to get wrong. Counts mislead, partial weeks look like crashes.
Compare, spot trends, report
0/3
Compare groups with rates instead of counts, read a trend without being fooled by partial weeks, and write the finding.
By the end of this stage
You can compare groups with rates, read a weekly trend without being fooled by short periods, and write a finding that includes a caveat.
How you know it worked
Finishing the lessons is not the goal. These are the things you should be able to do afterwards, and each one is worth checking honestly.
- Write any request as four parts: metric, rows, comparison and time range.
- Reconcile two tables and report the gap in rupees instead of picking one silently.
- Explain in one sentence why the mean and the median of the same column differ.
- Compare groups with a rate and show the number of attempts beside it.
- Produce a summary table from code and a three-line finding with a caveat.
Finish this track, get a certificate
Complete all 8 lessons and claim a certificate with a link you can add to LinkedIn or send to a recruiter.
Mistakes to avoid on this track
| Common mistake | What to do instead |
|---|---|
| Summing every order, including cancelled ones. | Filter to the status you mean (for revenue, paid) and say so in the report. |
| Comparing groups by count when the groups are different sizes. | Use a rate, and keep the attempts column next to it. |
| Reading the last, partial period as a decline. | Check the date range, then compare full periods or use a per-day figure. |
| Reporting a number without a caveat. | Add the sample size, the date range and one thing that could change the answer. |
Where to practise this
Where to go after this
Data Analysis Foundations reviews & rating
Finished a few lessons? Tell us what worked and what was confusing. Reviews are read by hand and help decide what gets rewritten next.