Unit 1 project: The mystery dataset

This unit you sorted apples, followed a trend and found hidden groups — one at a time, with the technique already chosen for you. Today nobody tells you which one. Your group opens a mystery dataset and decides.

Work in groups of 3–4. One period. Everything in mystery-<letter>.ods.

Each group gets one of these — you won't know which technique it needs until you look:

• Dataset P: 24 saffron flowers — petal length, colour score, and ready to pick: yes / no
• Dataset Q: 24 weeks — the week number and kilograms of apples a small orchard packed
• Dataset R: 24 songs — speed and loudness, with no names or labels at all

1. Look first. What are the features? Is there a label column? Is the answer a group, a number, or groups nobody named?
2. Choose the technique and write one sentence: "This is ___ because ___."
3. Keep some back — before anything else. Move the last 6 rows to a new sheet called Test. Don't look at them again until step 6.
4. Predict: how many of the 6 test rows will your model get right (or how close will it be)?
5. Build it on the other 18 rows:
   • classification → a rule, like Uncle's apple rule
   • regression → a scatter chart with a trend line
   • clustering → a scatter chart, clusters coloured and named by you
6. Test it on the 6 hidden rows. Count the right answers, or work out each error. For clustering, place each new row in a cluster and say whether it fits.

Each person writes two sentences: one thing the model does well, and one situation where it would go wrong. Sign with your roll number. This part is marked for you alone.

Each group says which dataset it had, which technique it chose, and its test result. Did the other groups with the same letter choose the same technique?