Stage 2 — Data Acquisition

An AI learns from examples, and nothing else. Whatever examples you hand it become its entire experience of the world. This stage is where you go and get them.

1What data acquisition is

Data Acquisition

Data acquisition is collecting the data your project needs, from reliable and authentic sources. This data becomes the base of the whole project.

The six stages, as a cycle

Tap a stage to stand on it, or press play and watch the work go round. Drag sideways to turn.

2. Data Acquisition — Collect the data the project needs, from reliable and authentic sources.

It answers: What data do we need, and where will we get it?

2Where data comes from

Source	Example
Surveys	Asking 200 farmers what they grow and what went wrong
Sensors and cameras	A camera in the field photographing leaves
Web scraping	Collecting published prices from websites
Open / government data	data.gov.in, census data, weather records
Records an organisation already keeps	A school's attendance and marks registers
Ready-made datasets	Datasets shared for learning, like Palmer Penguins

Reliable and authentic — both words matter

Reliable means the source is accurate and does not keep changing its mind. Authentic means it is genuine — it really is from who it says it is from. A government weather record is both. A number someone posted in a group chat is neither.

Cooking with what you were given

A cook can only cook with the ingredients brought into the kitchen. If the vegetables are rotten, no amount of skill saves the meal. Data is the ingredients, and every later stage is the cooking.

3Why bad data is so expensive

Everything after this stage trusts this stage. If you collect photos of healthy leaves taken only in bright sunshine, the model will decide that “healthy” means “bright” and call every evening photo diseased. It is not being stupid. It is telling you exactly what you showed it.

Two questions to ask about any dataset

Is there enough of it? Ten photos will not teach anything. Does it cover everybody? If your project is for the whole district, data from one village is not enough — and that gap is where bias comes from.

4Practice

A school wants to build an AI system that predicts which students are likely to fall behind. Suggest two reliable sources of data for it. (Practice)

2 marks

▶Show the answer

The school’s own attendance records and its past examination results. Both are kept by the school itself, so they are authentic, and both are recorded regularly, so they are reliable.

5Check yourself

Quick Check

Which stage of the AI Project Cycle is data acquisition?

Quick Check

Why must data come from authentic sources?

← Previous

🎯 Stage 1 — Problem Scoping

Stage 3 — Data Exploration 📊