Blogs · Data · Analytics

An Overview of Big Data Analytics

A concise overview of big data analytics: batch and streaming workloads, descriptive and predictive analysis, data quality, storage, compute, and communication.

2020.01.02 · 2 min read · by Zhenlin Wang

Introduction

Big data analytics is the practice of extracting useful information from data that is too large, fast, varied, or distributed for simple manual analysis.

The point is not that the data is “big.” The point is that ordinary tools and workflows are no longer enough.

Types of Analytics

Descriptive Analytics

Descriptive analytics explains what happened.

Examples:

Diagnostic Analytics

Diagnostic analytics explains why something happened.

Examples:

Predictive Analytics

Predictive analytics estimates what may happen next.

Examples:

Prescriptive Analytics

Prescriptive analytics recommends actions.

Examples:

Batch and Streaming

Batch analytics processes data in scheduled chunks. It is good for reports, backfills, training datasets, and daily metrics.

Streaming analytics processes events as they arrive. It is useful for monitoring, anomaly detection, event enrichment, and low-latency systems.

Many real systems use both: streaming for freshness and batch for correctness, reconciliation, or heavy processing.

Data Quality

Analytics quality depends on data quality.

Track:

Dashboards can look polished while being wrong. Build data quality checks into the pipeline.

Storage and Compute

Common storage patterns:

Common compute patterns:

Choose based on access patterns, not tool popularity.

Communicating Results

Analytics work is only useful when it changes understanding or action.

A good analysis states:

Avoid burying the conclusion under charts. Say what changed, why it matters, and how confident you are.

Closing

Big data analytics is an engineering and reasoning discipline. Reliable pipelines, clear metrics, data quality checks, and careful communication matter as much as the computation engine.