Why Panic Sets In the Moment I Ask About Their Notebook

Interview Insights · Data Analytics

Why Panic Sets In the Moment I Ask About Their Notebook

SECTION 01The three triggers I see behind notebook panic

When panic arrives, it tends to arrive at the same moment: right after I ask the candidate to share their screen and open the notebook. Before that, they talked about the project fluently. So the panic is not about the project idea. It is about the file.

In my experience it comes from three triggers. The first is time: the notebook was written long ago, and the details have faded. The second is size: a very long notebook with no headers feels like a maze when someone is watching. The third is guilt: parts of it were copied or adapted, and the candidate fears I will land on exactly that cell. Each trigger needs a different fix, so it is worth knowing which one is yours.

Interviewer: Can you share your screen and open
the notebook?

Candidate: Sure ... (opens file, scrolls fast)
... one second ... it is somewhere here ...
(scrolls back up) ... sorry, it is a long notebook.
I think the model part is ... here? No ...

(Forty seconds pass. The interviewer has seen
code, but heard nothing about the work.)

Interviewer: Can you share your screen and open
the notebook?

Candidate: Sure. It has four sections: loading,
cleaning, the price analysis and conclusions.
I will open the table of contents on the left so
we can jump around.

Shall I start with the price analysis in section 3?
That is where the main finding is.

The calm version contains no technical content at all. It simply proves the candidate knows the shape of the notebook. That alone stops the spiral, because the next step is always clear.

In Uncodemy mock interviews, 'share your screen and open the notebook' is one of the first requests we make, precisely because it shows so quickly whether a candidate knows the shape of their own work.

SECTION 02Re-familiarise yourself fast with a profile helper

If time distance is your trigger, the fastest cure is to meet your data again, not just your code. Reading cells top to bottom is slow and passive. Instead, run a small profile helper on the raw and cleaned DataFrames. In a few minutes you see the shape, column types, missing values and the main categories again, and the decisions in your cleaning cells suddenly make sense once more.

The helper below uses only standard pandas. Run it on the raw file and on your final cleaned table. The differences between the two outputs are your cleaning story: which columns disappeared, where missing values were filled and how many rows were removed. That story is exactly what an interviewer will ask about.

import pandas as pd

def profile(df, name, top=3):
    print(f'=== {name}: {df.shape[0]} rows, {df.shape[1]} columns ===')
    info = pd.DataFrame({
        'dtype': df.dtypes.astype(str),
        'missing': df.isna().sum(),
        'missing_pct': (df.isna().mean() * 100).round(1),
        'unique': df.nunique(),
    })
    print(info.sort_values('missing', ascending=False))
    for col in df.select_dtypes(include='object').columns:
        print(f'\n{col}: top values')
        print(df[col].value_counts().head(top))

raw = pd.read_csv('data/listings_raw.csv')
clean = pd.read_csv('data/listings_clean.csv')

profile(raw, 'RAW')
profile(clean, 'CLEAN')

AFTER RUNNING profile() ON RAW AND CLEAN:

Rows removed: ______ (raw) -> ______ (clean)
Reason(s): _________________________________

Columns dropped: ___________________________
Why: ______________________________________

Missing values handled:
  column ________ : filled with ____ because ____
  column ________ : dropped rows because ________

Key categories I should mention: ___________

Doing this the evening before an interview changes how you feel when the notebook opens. The cleaning cells stop being strangers. You recognise each one as the answer to a problem you just saw again in the profile.

SECTION 03A notebook map and recovery lines for the blank moment

Even well-prepared candidates sometimes blank out. The goal is not to never blank. It is to have a plan when it happens. Two tools make that plan concrete: a one-page notebook map, and a few honest recovery lines.

The notebook map is a short list of sections, the key cell in each, and one sentence about what it does. Keep it on paper beside your laptop during an online interview, or memorise it for an in-person one. Recovery lines are sentences you say out loud while you regain your footing, such as 'Let me use the search to find where I calculate the price per square foot.' They keep the interviewer with you while you find your place. Silence makes panic grow; a calm sentence shrinks it.

NOTEBOOK MAP: House price analysis (Noida listings)

1. Load            cell 2   read listings, parse dates
2. Clean           cell 5   drop listings with no area
                   cell 7   price per sq ft = price / area_sqft
                   cell 8   cap extreme price per sq ft at 1st/99th pct
3. Analysis        cell 12  median price per sq ft by sector
                   cell 14  bar chart, top 10 sectors
                   cell 16  price vs distance to metro scatter
4. Conclusions     cell 19  findings and limits

Variables to remember: listings, clean, by_sector
One-line finding: price per sq ft falls with metro distance.

JUPYTERLAB NAVIGATION UNDER PRESSURE

- Table of contents panel: left sidebar, lists every
  markdown header; click to jump.
- Search in notebook: Ctrl+F (Cmd+F on Mac).
- Collapse a long output: click the blue bar left of it.
- Run the current cell: Shift+Enter.
- See defined variables in a code cell:
      %whos

Practise each shortcut twice before the interview,
so your hands know them when your mind is busy.

The %whos magic lists the variables currently in memory with their types. It is a quiet lifesaver when you forget what you named your cleaned DataFrame, and using it confidently looks like expertise, not panic.

SECTION 04The night-before routine that prevents notebook panic

Follow this routine the evening before any interview where you might share a notebook. It targets all three triggers: time distance, size and borrowed sections.

NIGHT-BEFORE ROUTINE (about 60 minutes)

1. Restart kernel and run all. Fix any error.   (10 min)
2. Run profile() on raw and clean data.          (10 min)
3. Fill the notebook map on one page.            (15 min)
4. Mark borrowed or adapted cells honestly.     (5 min)
5. Practise opening, sharing and jumping to
   section 3 using the table of contents.        (5 min)
6. Say the one-line finding out loud 3 times.    (5 min)
7. Close the laptop. Sleep.

TEN MINUTES BEFORE THE CALL

[ ] Notebook open, outputs fresh, zoom increased
[ ] Table of contents panel open
[ ] Long outputs collapsed
[ ] Notebook map on paper beside the laptop
[ ] Other tabs and notifications closed
[ ] Glass of water nearby
[ ] One slow breath before you say 'Sure, sharing now'

WRITE THESE ON A CARD:

"Let me search for where I calculate ____."
"Let me check the data this step changes."
"That error is useful, let me read it."
"Can I take ten seconds to collect my thoughts?"
"That part was adapted from ____; the ____
 is mine, can I show you that?"

Read them aloud twice. Under pressure, the
sentences you have said before come out first.

SECTION 05Habits that keep notebook panic away

These habits make your notebooks easier to return to, which is the real long-term cure for panic.

• Write for your future self — Add a short markdown line under every header saying what the section does. In three months, that future reader will be you, under pressure.
• Keep notebooks short — Move exploratory work to a separate scratch notebook. A focused main notebook is far easier to navigate in front of an audience.
• Rerun before every interview — Stale outputs and hidden errors are common panic sources. A fresh run the night before removes surprises.
• Know your borrowed parts — Panic often hides guilt. Listing adapted cells and being ready to say so removes the fear of being found out.
• Narrate every scroll — Say where you are going before you scroll. It slows your hands down and keeps the interviewer following you.

SECTION 06Interview Q&A — Notebook panic

SECTION 07Test yourself — Notebook panic quiz

SECTION 08Frequently asked questions

SECTION 09Related reads