MBSA2113 Modern Data Management – Data Architecture Analysis

Question 1: Data Architecture & Strategic Decision-Making

A regional logistics company operating across Southeast Asia faces the following challenges:

Tasks:

(a) Design a modern data architecture suitable for this organization.

(b) Clearly justify your choice of data storage models (e.g., data lake, data warehouse, lakehouse).

(c) Explain how data flows from source systems to decision dashboards.

(d) Discuss how your architecture balances scalability, governance, and cost.

❖ Your answer must demonstrate how data “moves, transforms, and gains value” throughout the architecture.

Question 2: Data Quality, Governance & Trust

Tasks:

(a) Identify three root causes of low data trust in modern organizations.

(b) Design a data governance framework addressing ownership, quality, and accountability.

(c) Explain how metadata, lineage, and validation rules improve trust.

(d) Propose KPIs to measure data quality maturity over time.

❖ Your answer must link governance mechanisms directly to managerial confidence.

Question 3: Big Data Processing & Analytics Justification

Management asks whether to use traditional RDBMS, NoSQL, or distributed big data platforms.

Tasks:

(a) Classify each data type and its processing needs.

(b) Recommend suitable data technologies for each.

(c) Justify why a single database solution is insufficient.

(d) Explain how integration across platforms enables holistic insights.

❖ Focus on “why” decisions are made, not just “what” tools are used.

Experts Answer on Above Questions on Data Analytics

Data Architecture and strategic decision making

Modern Data Architecture – with respect to the Logistics Company, the most suitable modern Data Architecture is Lakehouse architecture. It’s main components include data sources like IoT sensors, mobile apps, ERP systems, data ingestion like Apache Kafka, data lake house such as databricks delta lake or snowflake, data processing using Apache spark, data warehouse layer and BI tools like Power BI/ tableau for dashboards.

Justification of data storage model

The lake house model is highly justified because it combines the benefits of data lake and data warehouse. It also supports structured and unstructured data, real time Analytics and eliminates any kind of data duplication.

Data flow to decision dashboards

The data generated from iot devices and mobile apps are ingested through streaming and batch pipelines. The raw data is stored in data Lake house and the cleaning of data is performed using spark. The curated data is stored in warehouse tables, and BI tools are utilised to access the data. The management finally receives real time dashboards.

Balancing scalability governance and cost

The factors that need to be considered are scalability which can be achieved through cloud based storage and distributed processing, governance through metadata management and cost using pay as you use cloud resources.

Data quality governance and trust

The three important causes of low data trust are poor quality of data, lack of data ownership and accountability, and inconsistency with respect to data across operational and analytical systems.

Data governance framework

With respect to ownership, it can be managed through assigning data owners and data stewards, quality by way of defining quality standards and validation rules, and accountability by establishing governance committees.

Role of metadata, lineage and validation rules

The metadata provides information about the source of data, and data lineage tracks the movement of data from source to dashboard, and validation rule helps in achieving accuracy and consistency.

KPI for data quality maturity

The important KPIs data accuracy rate, data completeness score, number of data quality incidents, time taken to resolve data issues, user satisfaction score.

Big data processing and analytics justification

Data like sales transactions can be classified as structured and need fast transactional processing, similarly customer reviews or classified as unstructured text and need text analytics for processing. Video analytics metadata is classified as semi structured and requires large scale storage and processing.

Suitable technologies

With respect to sales data, the technology in the form of relational databases like MySQL is appropriate whereas for customer reviews, NoSQL database is appropriate. For video metadata, Hadoop distributed platform is effective.

Why one data is insufficient

A single database is not effective because it is not capable of managing the structured transactions, unstructured text and large scale semi-structured data.

Integration across platform for Holistic insights

The integration of data from RDBMS, NoSQL and big data platforms is performed into a central analytics environment such as lake house and it allows for combining sales trends with customer sentiment, linking customer feedback to product performance.

Want Detailed Answers with References?

The analysis of data lakes, data warehouses, lakehouses, metadata, lineage, data quality, governance framework and analytics project above revealed significant findings and if you need a similar kind of assignment involving data analytics, get in touch with our professional data analytics experts in Malaysia. Simply visit the best assignment help in Malaysia to get quality assistance with your Data Analytics assignment.