Aim

To create a pandas Series object s5 containing numbers, compute the square of every element into a second Series s6 using vectorised arithmetic, and display only the squared values greater than 15 using boolean masking.

CO Mapping: CO1, CO2, CO4

Theory

A Series is a one-dimensional labelled array — the building block of every DataFrame column. Two ideas from this practical power almost all pandas analytics:

1. Vectorisation. s5 ** 2 applies the power operation to every element at once. Internally the loop runs in optimised C inside NumPy, so no Python for loop is written. Vectorised code is both faster (often 10–100×) and closer to mathematical notation, which reduces bugs.
2. Boolean masking. A comparison such as s6 > 15 does not return one True/False — it returns a Series of booleans, one per element, aligned by index. Passing that mask back into the Series (s6[s6 > 15]) keeps only the rows where the mask is True. This is the filtering primitive behind every "show me records where…" question in analytics.

A third, quieter idea is index preservation: filtering never renumbers the survivors. The labels of the kept elements travel with them, so you can always trace a filtered value back to its original observation.

Dataset

Procedure

1. Import pandas as pd.
2. Create s5 = pd.Series([1, 2, 3, 4, 5, 6]) — pandas assigns the index 0–5.
3. Compute s6 = s5 ** 2. No loop is written; the operation broadcasts across all six elements.
4. Print both Series with .to_list() for compact output.
5. Build the mask s6 > 15 and inspect it mentally: it is [False, False, False, True, True, True].
6. Apply the mask — s6[s6 > 15] — and print the surviving values.

Interpretation of Results

The filter returns 16, 25 and 36 — and, crucially, they keep their original index labels 3, 4 and 5. That tells the analyst which original observations survived, not just their values. Reading further: squaring is a monotonic transform for positive numbers, so the same three elements would be selected by the equivalent condition s5 > 3.87 (the square root of 15) — recognising such equivalences lets you filter on raw or derived columns interchangeably, whichever is cheaper or clearer.

Common Mistakes

1. Writing a Python loop (for x in s5: ...) instead of the vectorised s5 ** 2 — it works but abandons the entire point (and the speed) of pandas.
2. Using Python's and/or to combine masks — they raise ValueError on Series; use & and | with parentheses, e.g. s6[(s6 > 15) & (s6 < 30)].
3. Expecting the filtered result to be re-indexed from 0 — it keeps labels 3, 4, 5 unless you call .reset_index(drop=True).

🎯 Viva Questions

1. What is a Series? A one-dimensional labelled array; every DataFrame column is a Series.
2. Why is s5 2 faster than a loop?** The elementwise loop runs in compiled C inside NumPy instead of interpreted Python bytecode.
3. What type does s6 > 15 return? A boolean Series aligned to s6's index.
4. How do you combine two filter conditions? With & (and) / | (or), wrapping each condition in parentheses.
5. What happens to the index after filtering? It is preserved — survivors keep their original labels.
6. How would you renumber the filtered result from zero? s6[s6 > 15].reset_index(drop=True).