← Learning Path / Step 2
Step 3 of 10 • ~2 min read
Gather a Tiny Dataset
Small is kind. A compact slice of data lowers the barrier and builds confidence.
📚 THEORY
A tiny dataset lowers the barrier to starting and makes the work feel human-sized.
If you try to begin your storytelling journey on a massive warehouse table with thousands of rows and dozens of columns, you quickly get overwhelmed and lose the narrative thread.
A compact slice of data — one team, a few months, a single process — lets you see patterns with your own eyes.
With simpler datasets, you can scroll the full table, understand each column, and mentally connect the numbers to real-world events. That builds intuition, which is more important at this stage than complex technique.
A small dataset is also safer for experimentation — you can try different views, filters, sort, and play without fear.
This playful exploration builds confidence and curiosity, two essential ingredients for long-term growth in data literacy. Crucially, a tiny dataset is enough to tell a meaningful story as long as it’s tied to a real question and decision.
Resist the temptation to gather everything first. More data is not more story. Start with the minimum dataset that could answer your question, then expand only if necessary.
“Not everything that can be counted counts, and not everything that counts can be counted.”
William Bruce Cameron
✏️ AISHA’S APPLICATION – STEP 3 IN PRACTICE
Aisha resists the urge to pull the full HR attendance database. Instead, she scopes her dataset to exactly what she needs:
- 5 sessions before the change (10:00 am)
- 5 sessions after the change (16:00)
She chooses just these columns:
- Time Related: Month, Date, Time Slot
- Capacity and Engagement: Seats Available, Sign-ups, Show-ups
- Quality: Average Feedback
This makes it easy for her and her audience to follow. Participants can recreate this dataset in Excel/Sheets and visually scan it to notice:
- Sign-ups are dropping from the low 20s to the low teens
- Feedback scores are staying around 4.5-4.7
You can explicitly ask:
“What’s the earliest month where you see a change? What stayed stable?”
Want to follow along with Aisha’s exact data? Download her dataset here.
💡 Try it now: For your question from Step 1, identify the smallest dataset that could answer it. Write down: which table, which columns, and what date range you need.