Data and an Introduction to Big Data
AdvancedData are symbols recording facts; the statistical process runs collect, organise, analyse, present. Big data — vast in volume, swift in arrival, rich in variety — reveals patterns for traffic control, epidemic tracking and shopping suggestions. Beware its traps: correlation is not causation, samples can be biased, charts can mislead; personal data are private, to be guarded and respected.
- Statistics runs collect, organise, analyse, present; bar, pie and line charts each serve their purpose.
- Big data is vast, fast and varied; its worth lies in the patterns and trends analysis reveals.
- Big data makes cities smarter: signal control, epidemic tracking, recommendations and planning.
文體小知識：數據科普說明文用生活例子解釋數據怎樣收集、整理和分析，並提醒讀者小心數據的陷阱。
§Text Passage
1The class board carries a bar chart: favourite sport in class — football 18 votes, basketball 11, swimming 9. That is data in daily life. A survey runs four steps: collect (design a questionnaire or tally), organise (sort and tabulate), analyse (counts, averages, trends), present (choose the right chart). Compare quantities with a bar chart, show shares of a whole with a pie chart, show change over time with a line chart; the right chart carries the message accurately.
2When data grow vast enough, we call them big data, with three marks: volume — terabytes even petabytes; velocity — tens of thousands of records a second; variety — text, images, locations, sensor readings. Hong Kong's Octopus logs millions of transactions daily; social posts and likes, the territory-wide weather sensor network — all feed big data. One record says little, but analysed in bulk the patterns and trends emerge.
3Big data is already changing city life. Transport: the Transport Department reads vehicle speeds from thousands of road sensors to adjust signals in real time and ease jams. Health: public health teams analyse consultation data to track epidemics early. Commerce: shopping platforms study browsing and purchases to recommend what you may like. Planning: crowd heat-maps guide footpath and resting-space design. Data make the city smarter and services kinder.
4Yet data can deceive, so build data literacy. First, correlation is not causation: ice-cream sales and drownings rise together in summer — not ice cream causing drowning but the shared cause of hot weather. Second, samples can be biased: surveying only the library society cannot speak for the whole school's tastes. Third, charts can mislead: crop the axis and a 2-per-cent gap looks like threefold. Finally, personal data are private, protected by the Personal Data (Privacy) Ordinance: a torch app asking for your contacts deserves refusal. Data never lie, but selective use can mislead — asking questions and verifying is how you master the data age.
1Summary 重點摘要
Data are symbols recording facts; the statistical process runs collect, organise, analyse, present. Big data — vast in volume, swift in arrival, rich in variety — reveals patterns for traffic control, epidemic tracking and shopping suggestions. Beware its traps: correlation is not causation, samples can be biased, charts can mislead; personal data are private, to be guarded and respected.
2Vocabulary cards Vocabulary
Symbols, numbers and records of facts which, organised and analysed, become useful information.
The methods of collecting, organising, analysing and presenting data.
Collecting views and facts from a group through designed questions.
A representative value: the sum of a set divided by its count.
The overall direction of change data show over time.
Data so vast, fast and varied that computers must analyse them for patterns.
The part drawn from a whole for study; it must represent the whole or conclusions skew.
Two sets of data rising and falling together — which need not mean one causes the other.
Systematic unfairness in collecting or presenting data, skewing conclusions from fact.
The right, protected by law, of personal data to be used only as their owner decides.
3Step by step Steps
- Step 1: Walk the four statistical stepsThe full statistical process runs in four steps — collect, organise, analyse, present — and the order matters. Example: surveying the class's favourite sport — design the options and gather votes (collect), tally each option (organise), count each (analyse), chart it (present). Each step lays the base for the next; once the order slips, data go missing or double up and the conclusion cannot be trusted.
- Step 2: Choose the right chartCharts are the language of data; choosing the right one lets readers see the point at a glance. Example: comparing four sports' votes needs a bar chart; shares of the class a pie chart; votes per day over a week a line chart — the wrong chart misleads readers about what the data says. Before choosing, ask whether you mean comparison, share or trend, and the answer becomes clear.
- Step 3: Know big data's three VsBig data has three features — volume, velocity and variety, the three Vs. Example: millions of Octopus transactions a day is volume; recorded instantly is velocity; spanning transport, shopping and access is variety. Only with all three can data show its full value; a single record says little, but bulk data analysed reveal patterns and trends.
- Step 4: Read big data at workThe value of big-data analysis lies in turning vast numbers into concrete action. Example: from thousands of speed sensors, the Transport Department lengthens green lights at jammed junctions by seconds — bulk data plus analysis lets a city tune itself. Similar methods serve weather forecasts, epidemic tracking and shopping recommendations; the principle is always to find patterns in mass records and turn them into decisions.
- Step 5: Separate correlation from causationTwo numbers rising together does not mean one causes the other — the commonest trap in reading data. Example: ice-cream sales and drownings rise together — not dangerous ice cream but the shared cause of hot weather; find the common cause before concluding. Verify first and judge later: that is where data literacy starts; behind any surprising link, first ask whether a third factor is at work.
- Step 6: Check the sample's fairnessHow far a conclusion can be trusted depends on whether the sample is fair and large enough — the sampling method matters more than the mere count. Example: asking only the library society about the whole school's book tastes is sample bias; draw randomly from every class — bigger samples, firmer conclusions. An unfair sample voids even the prettiest result; when an advert claims 'eight in ten users are satisfied', first ask where the sample came from.
- Step 7: Guard personal dataIn the data age everyone must learn to guard personal information and respect the privacy of others. Example: a torch app demanding contacts — refuse; answer only necessary questionnaire items; never post others' grades or addresses — privacy starts with you. Such habits cut the risks sharply; personal data are protected by the Personal Data (Privacy) Ordinance, and before giving anything, ask whether it is truly necessary.
4Diagrams Diagrams
5Worked example Example
Example: Yan wants to study whether classmates' daily screen time relates to PE test results. Write how to collect the data, two fairness principles to observe, and one caution when interpreting.
- Collection: an anonymous questionnaire recording daily screen time (under one hour, one to two hours, over two hours) and the grade from one and the same PE test, with all forty classmates joining.
- Fairness one: anonymity — pupils answer truthfully about screen time without fear of blame, cutting under-reporting.
- Fairness two: same-term comparison — every group uses the same test's grades; mixing in last term's results would invalidate the study.
- Caution: even if the longest-screen group averages lower, she may say only 'there is a correlation', not 'screen time causes lower grades' — sleep, exercise habits and other factors may act; suggest interviews to probe the real reasons.
- Answer: Yan should gather screen time and grades together anonymously, compare like with like, and read the result as correlation only, mindful of other factors — that is the heart of data literacy.
6Common mistakes Mistakes
7Quick check Quick check
Q1What are the four steps of a statistical survey?
Q2What do the three Vs of big data mean?
Q3Give one example of big data improving city life.
Q4Distinguish correlation from causation with an example.
Q5Name two ways to protect personal data.
8Recap Recap
- Statistics runs collect, organise, analyse, present; bar, pie and line charts each serve their purpose.
- Big data is vast, fast and varied; its worth lies in the patterns and trends analysis reveals.
- Big data makes cities smarter: signal control, epidemic tracking, recommendations and planning.
- Three cautions of data literacy: correlation is not causation, samples must represent, charts can mislead.
- Personal data are private: check permissions, share carefully, and respect everyone's data including your own.