Data Acquisition

Structured Data

Reason — Structured data is neatly organised and easily searchable, often found in databases or spreadsheets. It is arranged in rows and columns, making it straightforward to input, store, query and analyse.

Unstructured Data

Reason — Unstructured data is information that does not have a predefined organisational format. It includes written documents, pictures, audio recordings and videos.

Semi-structured Data

Reason — Semi-structured data blends features of both structured and unstructured data. For example, an email has structured components such as the sender, recipient, subject line and date, along with unstructured content in the message body.

Quantitative Data

Reason — Quantitative data involves numerical values and measurements. It quantifies characteristics, behaviours or other specified variables, making it possible to generalise findings from a broader sample.

To create new variables from raw data to enhance model performance

Reason — Feature engineering enhances AI models by creating new variables from raw data, which can help the models to better understand and predict outcomes.

Web Scraping

Reason — Web scraping involves using a program to automatically collect data from websites.

1. Structured data is neatly organised and easily searchable, often found in databases or spreadsheets.
2. Information that does not have a predefined organisational format is known as Unstructured data.
3. Semi-structured data blends features of both structured and unstructured data.
4. Descriptive information collected through interviews, surveys or observations is called Qualitative data.
5. Quantitative data involves numerical values and measurements that quantify characteristics, behaviours or variables.
6. To identify data requirements for an AI project, one must first identify different users of the project.
7. Feature engineering involves creating new variables from raw data to enhance the functionality of AI models.
8. A visual representation showing the components, processes and interactions within a system is called System map.

Two examples of structured data:

1. School attendance register
2. Spreadsheet containing students' records

Two sources of data acquisition:

1. Databases
2. APIs (Application Programming Interfaces)

Two types of data in AI projects:

1. Structured Data
2. Unstructured Data

Two steps in identifying data requirements:

1. Identify different users of the project.
2. Identify the exact task they will be performing through the application.

Two components of a system map:

1. Entities
2. Interactions

Both A and R are true and R is the correct explanation of A.

Reason — Unstructured data does not have a predefined organisational format, making it more difficult to process and analyse. Since this data lacks a uniform structure, advanced analytical tools and techniques are required to interpret and extract meaningful information from it.

(a) Structured data

(b) Unstructured data

Reason — Unstructured data is information that does not have a predefined organisational format. It includes written documents, pictures, audio recordings and videos.

(c) Semi-structured data

Reason — Semi-structured data blends feature of both structured and unstructured data. For example, an email has structured components such as the sender, recipient, subject line and date, along with unstructured content in the message body.

Structured Data is neatly organised and easily searchable, often found in databases or spreadsheets. This type of data is arranged in rows and columns, making it straightforward to input, store, query and analyse. For example, a school’s attendance register is structured data, with rows for individual entries and columns for attributes like names, roll numbers and attendance. The organisation of structured data allows for efficient data retrieval and management.

Unstructured Data is an information that does not have a predefined organisational format, making it more difficult to process and analyse. This category includes various formats like written documents, pictures, audio recordings and videos. Since this data lacks a uniform structure, advanced analytical tools and techniques are necessary to interpret and extract meaningful information from it.

Semi-structured Data blends feature of both structured and unstructured data. It does not conform perfectly to traditional databases but contains some organisational elements that make it easier to handle than purely unstructured data. For example, an email has structured components such as the sender, recipient, subject line and date, along with unstructured content in the message body.

Qualitative Data represents descriptive information and is usually collected through interviews, surveys or observations. This type of data is non-numerical and offers insights into behaviours, preferences or opinions. It captures subjective elements and helps in understanding underlying motives and perspectives.

Quantitative Data involves numerical values and measurements. It quantifies characteristics, behaviours or other specified variables, making it possible to generalise findings from a broader sample. This data type is vital for statistical analysis, enabling well-informed decisions based on numerical insights.

System Map is a visual representation that shows the components, processes and interactions within a system. It helps in understanding how different parts of a system connect and interact with each other. In the context of AI, system maps are useful for visualising data flow, data processing steps and the roles of various entities in the system.

In an AI project, both qualitative and quantitative data are used to understand the problem better and improve system performance.

Qualitative data represents descriptive information and is usually collected through interviews, surveys or observations. It is non-numerical and helps in understanding behaviours, preferences and opinions. For example, in an AI library assistant project, students’ feedback about their favourite books or their opinions on reading preferences can be used to understand user interests and improve book recommendations.

Quantitative data involves numerical values and measurements. It quantifies characteristics, behaviours or variables and is useful for statistical analysis. For example, data such as the number of books issued, frequency of book borrowing, students age or total number of users can be analysed by the AI system to identify patterns and make well-informed decisions.

Thus, qualitative data provides insights into user preferences, while quantitative data supports analysis and decision-making in AI projects.

Identifying stakeholders in an AI project is significant because it ensures that the project meets the needs and expectations of those who are involved in or affected by the project. By understanding the requirements and roles of stakeholders, the AI system can be designed more effectively and aligned with its intended purpose, leading to better outcomes and successful implementation.

The process of creating a system map for an AI project involves the following steps:

1. Define the Scope: Decide which part of the system to map, such as a specific process, data flow or the entire system.
2. Identify Entities: List all components involved, such as databases, sensors, software applications and users.
3. Map Interactions: Draw lines to show how entities are connected and interact with each other.
4. Detail Processes: Outline the key processes occurring within the system.
5. Show Data Flows: Indicate how data moves through the system from collection to utilisation.

Structured data is neatly organised in rows and columns, making it easy to input, store, query and analyse. Unstructured data does not have a predefined organisational format, making it more difficult to process and analyse. Semi-structured data blends features of both; it does not fit perfectly into traditional databases but contains some organisational elements, making it easier to handle than purely unstructured data.

Semi-structured data combines elements of both structured and unstructured data. Although it does not conform perfectly to traditional databases, it contains some organisational elements that make it easier to handle and manage than purely unstructured data. This allows for more efficient data management and analysis.

Surveys and questionnaires are tools used to collect data directly from people by asking them questions. They help AI projects gather information about user preferences, opinions and habits.

In an AI project, they can be used effectively to understand user needs and personalise system outputs. For example, in an AI library assistant project, surveys can collect data on students' favourite genres, reading habits and book preferences. This helps the AI system make personalised and relevant book recommendations.

Ethical issues related to privacy and consent in data collection arise because data often includes personal and sensitive information. Therefore, users should be informed about what data is being collected, how it will be used and for what purpose. Their consent should be obtained before collecting or using their data. The collected data should also be stored securely and protected from unauthorised access. These steps help prevent misuse of data and ensure responsible use of data in AI projects.

Data quality assurance is important in AI projects because the accuracy, reliability and performance of AI models depend directly on the quality of data used. Accurate, comprehensive and relevant data ensures that AI systems produce meaningful results and make correct decisions. Poor-quality data can lead to errors, biased outcomes and unreliable models. To ensure data quality, data should be collected from reliable sources, checked for errors and inconsistencies, and cleaned before use. Irrelevant or duplicate data should be removed, and data should be regularly reviewed and updated to maintain its accuracy and completeness. These methods help in building trustworthy and effective AI projects.

Sensors and Internet of Things (IoT) devices are tools that collect data from the physical environment. They help AI projects by continuously gathering real-time data that can be analysed and used for decision-making. For example, in a smart library, sensors can be used to monitor book movement, such as which books are taken off the shelves most frequently, and environmental conditions like temperature, lighting and noise levels. This data helps the AI system understand usage patterns and maintain a comfortable environment, thereby improving the effectiveness of the AI project.

Web scraping is a process used to automatically collect data from websites using a computer program. In data acquisition for AI projects, web scraping helps gather large amounts of useful information from online sources efficiently. For example, web scraping can be used to collect book descriptions, summaries, author information and reader reviews from online bookstores or publishers websites. This allows an AI system to access a wider range of data, making it more informative and useful for analysis and decision-making.

A well-defined goal helps an AI project by providing clear direction and focus. It helps the project team understand what needs to be achieved, avoids confusion and ensures that efforts are aligned with the expected outcomes. Clear goals also make it easier to evaluate the performance of the AI system.

The steps involved in setting goals for an AI project include identifying the problem to be solved, defining clear and specific objectives, understanding the needs of users and stakeholders, deciding the scope of the project and determining success criteria.