Step 1 — Data Collection

Phase: Data Analysis

The Conversation in the Real World

“Before we build anything, we need to make sure we actually have the data — and that it’s the right data.”

In a real marketing scenario, this step is where you’d ask:

• Where does this data come from? (GIA grading reports, point-of-sale records, competitor scraping)
• Is it complete? Do we have enough rows to train a reliable model?
• Does it cover the range of diamonds we actually sell?

   

Diamond Dataset Columns:

•  

  Column	Type	Range / Values	Description
  price	int	$326 – $18,823	Target variable — price in USD
  carat	float	0.2 – 5.01	Diamond weight
  cut	category	Fair, Good, Very Good, Premium, Ideal	Quality of the cut
  color	category	D (best) to J (worst)	Diamond colour grade
  clarity	category	IF (best) to I1 (worst)	Clarity measurement
  depth	float	43 – 79	Total depth percentage
  table	float	43 – 95	Top width relative to widest point
  x	float	0 – 10.74 mm	Length
  y	float	0 – 58.9 mm	Width
  z	float	0 – 31.8 mm	Depth

   

  💡 Marketing Manager’s Perspective: Think of this table as your product catalog. Each row is one diamond on your shelf. The price column is what you charged (or would charge). Everything else describes why that diamond is worth that price.