What's changed: Created Cloud Digital Leader Chapter 2 (Domain 2 "Data": value and types of data = structured/semi-structured/unstructured, batch/streaming; Google Cloud data management = Cloud Storage/Cloud SQL/Spanner/Firestore/Bigtable/BigQuery/Dataflow/Pub/Sub/Looker).
2.1The value of data and types of data
Understand the value of data as fuel for business decisions and AI, the difference between structured, semi-structured, and unstructured data, batch vs streaming processing, and the data lifecycle from generation to value.
For modern organizations, data is one of the most important assets. Analyzing accumulated data leads to better decisions and new services, and it is also the "fuel" for AI/machine learning. Yet data creates no value by itself—it gains meaning only when collected, prepared, and analyzed. First, understand the types of data and how it is processed.
2.1.1Three types of data
| Type | Characteristic | Example |
|---|---|---|
| Structured | Fixed rows/columns (schema) | Sales tables, customer tables |
| Semi-structured | Loosely structured via tags/keys | JSON, XML |
| Unstructured | No fixed format | Images, video, text, audio |
Structured data has a fixed row/column format like spreadsheets or relational databases, making it easy to query and aggregate. Semi-structured data like JSON or XML is loosely structured via tags/keys. Unstructured data like images, video, text, and audio has no fixed format and makes up the majority of the world’s data. Recently AI has made unstructured data usable, increasing its value.
2.1.2Batch vs streaming processing
There are two main ways to process data. Batch processing accumulates a volume and processes it together (e.g., aggregating a day of sales overnight). Streaming processing processes data in real time as it arrives (e.g., anomaly detection for sensors or payments). Choose by when you need results: streaming when immediacy matters, batch when periodic is enough.
Common: example → type/processing. E.g., "images/video/audio" = unstructured; "JSON/XML" = semi-structured; "row/column sales table" = structured; "real-time anomaly detection as data arrives" = streaming; "aggregate overnight in bulk" = batch; "data is fuel for AI" = value of data.
Watch the mix-ups: (1) structured (fixed format) / semi-structured (JSON etc.) / unstructured (images, text). (2) batch (accumulate then process) vs streaming (real time)—choose by whether immediacy is needed. (3) Collected data is not value until prepared and analyzed.
2.1.3Section summary
- Data fuels decisions and AI; it becomes value only when prepared and analyzed
- Types = structured (tables) / semi-structured (JSON, XML) / unstructured (images, video, text, audio)
- Processing = batch (accumulate) vs streaming (real time); choose by immediacy
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. Which data has no fixed format, like images, video, audio, and free-form text?
Q2. Which data has a fixed row/column format and is easy to query and aggregate?
Q3. Which data is loosely structured via tags or keys, like JSON or XML?
Q4. Which processing suits detecting anomalies in real time as sensor or payment data arrives?
Q5. Which processing aggregates a day of sales together overnight?
Q6. Which best describes the value of data?