Facets of Data

Chapter:4

Descriptive Analytics

Facets of Data

•  The word facet means side or aspect of something. Hence the term Facets of Data means ‒ different sides or forms in which data can exist or be represented. For instance ‒ Data can be represented in tabular form, text, multimedia, real‒time and so on.

•  Following are different forms in which data appears ‒

1. Structured data: Data is well organized in rows and columns ‒ like Excel sheet or database table. Every column has fixed data type like name, age, marks etc.

2. Unstructured data: Data is without any fixed format. There is no predefined structure for data. It is in free‒form.

For example ‒ Social Media Post, Chat messages or emails.

3. Natural language data: Data is written in human language such as English, Hindi and so on. But machine find it hard to interpret directly. For example ‒ "This product is good" is a review text. Or "Google Play music" is a voice command. This data is considered as a natural language data. Natural Language processing is a part of AI that helps computers understand human language.

4. Machine generated data: Data is automatically created by devices, sensors or systems. It is not generated by humans. For example ‒ Temperature sensors data in weather station, smartwatch heart rate logs.

5. Graph based data: This is a kind of data in which data is represented as nodes(entities) and edges(relationships). It shows how the things are connected. For instance ‒ Facebook showing friends and their connections. Google maps showing cities and routes.

6. Audio, video, image data: This is a multimedia data. It is the data that contains sound, visuals or both. This data is complex and rich in information. For instance ‒ Photos on Instagram or YouTube Videos.

Streaming data: This data is the data that comes continuously in real‒time, rather than being stored first. For instance Stock market price updates, social media live comments.

Importance of considering facets of data

1. Right data storage method: Each type of data needs a different storage system. For example ‒ Structured data can be stored in relational databases unstructured data can be stored in NoSQL databases like MongoDB, Cassandra.

2. Analysis technique : Different data forms require different analysis methods. For instance to analyse tweets we can't use Excel we need text mining or NLP tools.

a. Structured data uses statistical methods.

b. Unstructured text uses Natural Language Processing.

c. Images / Audio and videos use computer vision.

d. Streaming data uses Real‒time analytics.

3. Data preprocessing: Understanding the facet helps to clean and prepare data correctly.

a. Structured data remove missing values, standardize columns.

b. Unstructured data extract keywords or features.

c. Audio and video data is converted to numerical features before analysis.

4. Selection of tools and techniques: Knowledge of facets help in selecting suitable tools and technologies.

a. If structured data is there then SQL, Excel, R or Python language is selected for processing.

b. If Unstructured data is there then Hadoop, Spark or NoSQL technologies are preferred.

c. If there is Image or Video data is present OpenCV, TensorFlow is used.

5. Efficient data integration: In real‒world applications, data rarely comes from one source. We often combine structured, unstructured and streaming data. For example ‒ An E‒commerce company uses structured data(product database), unstructured data(customer reviews) or streaming(real‒time website clicks).

6. Decision making: Different data forms give different kinds of insights. For instance ‒ Structured data tells us what are the contents, unstructured data tells why did it happen, and streaming tells use what is happening right now.

7. Data governance and security: Each type of data has different privacy, storage and access methods. For example ‒ Text or Emails may need encryption. Knowing the facet helps design correct security policies.

Review Questions

1. What does "Facets of Data" mean? Explain any three forms of data.

2.What is the importance of considering facets of data.