Why data analytics matters¶

Every organisation makes decisions. Data analytics is the discipline of making those decisions with evidence instead of guesswork, and of explaining that evidence so other people act on it.

Before we write any code or build any charts, this Foundations section answers a simple question: why bother? You will see how analysis has changed real outcomes, how numbers can mislead, and why the way we present evidence matters as much as the evidence itself.

---

What is data analytics?¶

Data analytics turns raw records into answers to business questions. It usually works at four levels:

Level	Question it answers	Everyday example
Descriptive	What happened?	Sales fell 8% last quarter.
Diagnostic	Why did it happen?	The fall came almost entirely from one region after a competitor opened.
Predictive	What is likely to happen next?	If nothing changes, the region will lose a further 5% next quarter.
Prescriptive	What should we do about it?	Test a local loyalty offer in that region before changing national prices.

This course focuses mostly on the first two levels, done properly: a clear question, trustworthy data, correct calculations, honest charts and a recommendation a manager can act on.

---

Real stories where data changed the outcome¶

A map that changed how we think about disease (London, 1854)¶

During a cholera outbreak in Soho, physician John Snow investigated where the victims lived and where they drew their water. The cases clustered around one public water pump on Broad Street. At the time, most experts blamed "bad air". Snow persuaded the local authorities to remove the pump's handle. His famous dot map of the deaths, drawn shortly afterwards, became the classic demonstration that cholera spreads through water, though it took years for that view to be widely accepted.

A chart that reformed hospitals (1850s)¶

After the Crimean War, Florence Nightingale showed with her "rose" diagrams that far more soldiers died of preventable disease in hospitals than of battle wounds. Her charts were designed for politicians, not statisticians, and helped drive major sanitary reforms.

Winning on a small budget (baseball, 2002)¶

The Oakland Athletics, with one of the lowest payrolls in Major League Baseball, used statistics such as on-base percentage to find undervalued players. They competed with teams spending far more. The story became the book and film Moneyball, and data-driven scouting spread across sport.

Preparing for a hurricane (retail, 2004)¶

Before a hurricane reached Florida, Walmart analysts looked at what shoppers had bought before earlier storms. Alongside torches and water, sales of strawberry Pop-Tarts and beer reportedly jumped well above normal. Stores were stocked accordingly.

When big data got it wrong (public health, 2008–2015)¶

Google Flu Trends tried to estimate flu levels from search queries. It worked for a while, then missed the 2009 swine-flu pandemic and badly overestimated later flu seasons, and was shut down. People's search habits and Google's own search algorithms changed, and the model was not re-checked against official flu data often enough.

---

Where you meet analytics every day¶

• Banking: a card payment is flagged as possible fraud because it does not fit your normal pattern.
• Streaming and shopping: recommendations are based on what similar customers watched or bought.
• Transport: ride-hailing prices rise when demand outstrips available drivers.
• Healthcare: hospitals plan staff rotas using predicted admission levels.
• Workplaces: HR teams track hiring time and staff turnover to spot problems early.

---

Why presentation is part of analysis¶

An analysis that nobody understands has no effect. The most important finding in a report can be lost in a cluttered table, a misleading chart or a slide with too much text.

Good analysts therefore do three jobs:

1. Find what is true in the data.
2. Check that it really is true.
3. Show it so the right person understands it quickly and acts on it.

The rest of this Foundations section covers each of these.

---

What comes next¶

Work through these pages in order:

1. Statistics are not enough: why you must always look at the data, not just the summary numbers.
2. How statistics can mislead: correlation versus causation, misleading charts and other traps.
3. Types of analysis: looking at one, two or many variables at once.
4. Designing clear visuals: data-ink, how the eye groups things, and using colour with purpose.
5. Storytelling with data: the Challenger disaster, and how to turn analysis into a message people act on.