📚 A-Level Mathematics: Key Points for Analysing Large Data Sets | A-Level 数学：大数据集分析要点
The Large Data Set (LDS) is a compulsory context in A-Level Mathematics. Exam questions rarely ask you to memorise the data, but they do expect you to understand its structure, sampling method, variables and limitations. This article summarises the essential statistical skills you need to analyse large data sets with confidence in your exams.
大数据集（LDS）是 A-Level 数学考试中的必考情境。考试很少要求你背诵数据本身，但确实期望你理解数据集的背景结构、抽样方法、变量及其局限性。本文系统梳理分析大数据集所需的核心统计技能，帮助你在考试中自信应对。
1. What Is the Large Data Set? | 什么是大数据集？
In the Edexcel specification, the Large Data Set is the Met Office's weather data, covering a range of locations in the UK and internationally. It contains daily observations such as daily mean temperature, daily total rainfall, daily total sunshine and daily mean windspeed. You should know the locations, the time period covered and the units used.
在 Edexcel 考纲中，大数据集来自英国气象局的天气数据，覆盖英国及全球多个地点。数据包含每日平均气温、每日总降雨量、每日总日照时数以及每日平均风速等观测值。你需要知道包含哪些地点、覆盖的时间段以及使用的单位。
Key locations include Camborne, Heathrow, Hurn, Leeming and Lerwick in the UK, plus Beijing, Perth and Jacksonville internationally. The Edexcel LDS covers observations from 1987 to 2015, although the exact date range may vary by exam series. Always read the preamble to the question carefully.
关键地点包括英国境内的坎伯恩、希思罗、赫恩、利明和勒威克，以及国际城市北京、珀斯和杰克逊维尔。Edexcel 大数据集覆盖 1987 至 2015 年的观测记录，但具体日期范围可能因考试系列而略有不同。务必仔细阅读题目开头的情境说明。
2. Sampling Techniques | 抽样方法
Large data sets are often too big to analyse in full, so statisticians take a sample. A-level questions may ask you to identify a sampling method and explain how to apply it, or to comment on the representativeness of a sample.
大数据集往往过于庞大，难以全部分析，因此统计学家会抽取样本。A-Level 题目可能要求你识别抽样方法、说明如何实施，或评价样本的代表性。
- Simple random sampling: every member of the population has an equal chance of being selected. Use lottery methods or random number generators.
- 简单随机抽样：总体中每个个体被选中的概率相等。可使用抽签法或随机数生成器。
- Systematic sampling: select every n-th member after a random starting point.
- 系统抽样：在随机起点之后，每隔 n 个个体选取一个。
- Stratified sampling: divide the population into strata, then sample proportionally from each stratum.
- 分层抽样：先将总体划分为若干层，再按比例从各层中抽样。
- Quota sampling: interviewers choose a fixed quota from each group, usually for people-based surveys.
- 配额抽样：调查者从每个组别中按固定配额选取个体，通常用于人群调查。
For the Met Office LDS, weather stations are not randomly located; they are fixed at specific geographical positions. This means the data set is a convenience or opportunity sample of weather conditions, not a random sample. In an exam, you should be able to state that the sample is biased towards the locations where weather stations exist.
对于气象局大数据集，气象站并非随机布置，而是固定在特定地理位置。因此，该数据集属于天气状况的便利样本或机会样本，而非随机样本。在考试中，你需要能够指出：该样本存在偏差，因为它只覆盖了设有气象站的地点。
3. Cleaning Data and Data Types | 数据清洗与数据类型
Real-world data often contains errors, missing values or impossible entries. Cleaning data means identifying and correcting these problems before analysis. Exam questions may show a small table of data and ask which entries are anomalous or missing.
真实世界的数据通常包含错误、缺失值或不合理的条目。数据清洗是指在分析之前识别并更正这些问题。考试题目可能给出一个数据表，要求你找出异常或缺失的条目。
- Data validation: checking data at the time of entry, e.g. a negative rainfall value should be rejected.
- 数据有效性检查：在录入数据时进行检查，例如降雨量出现负值应被拒绝。
- Missing data: if a value is missing, you might estimate it, omit the case, or note the gap as a limitation.
- 缺失数据：如果某个值缺失，你可以估计它、删除该记录，或将缺口列为局限性。
- Units: check whether measurements are in °C, mm, hours and knots, and convert if needed.
- 单位：检查测量值是否使用 °C、mm、小时和节等单位，必要时进行换算。
Data are also classified by type. Categorical data can be nominal (e.g. location name) or ordinal (e.g. wind strength: light, moderate, strong). Numerical data can be discrete (e.g. number of rain days) or continuous (e.g. temperature). The type of data determines which charts and statistical measures are appropriate.
数据还可以按类型分类。分类数据可以是名义数据（如地点名称）或顺序数据（如风力等级：轻、中、强）。数值数据可以是离散数据（如下雨天数）或连续数据（如气温）。数据类型决定了应使用哪些图表和统计量。
4. Measures of Central Tendency | 集中趋势度量
The mean, median and mode summarise the 'typical' value of a data set. In large data set questions, you may use a calculator to find the mean and median of a sample, or compare two locations using these measures.
平均数、中位数和众数用于概括数据集的“典型”取值。在大数据集题目中，你可能需要使用计算器计算样本的平均数和中位数，或比较两个地点的集中趋势。
For a data set with n values, the sample mean is given by:
x̄ = (Σx) / n
where Σx is the sum of all data values. The median is the middle value when data are arranged in order. If n is odd, the median is the ((n + 1) / 2)-th value. If n is even, it is the midpoint of the two middle values.
其中 Σx 是所有数据值的总和。中位数是将数据按顺序排列后的中间值。若 n 为奇数，中位数是第 ((n + 1) / 2) 个值；若 n 为偶数，则取中间两个值的平均数。
The mode is the value that occurs most frequently. For continuous data, we often give a modal class or modal group. Remember: the mean is affected by extreme values, but the median is not.
众数是出现次数最多的值。对于连续数据，我们通常给出众数区间。请记住：平均数受极端值影响，而中位数不受极端值影响。
5. Measures of Spread | 离散程度度量
Measures of spread show how consistent or variable the data are. The simplest measure is the range, but it is easily distorted by outliers. The standard deviation and interquartile range are far more useful for large data sets.
离散程度度量反映数据的一致性或波动幅度。最简单的度量是极差，但它容易被异常值扭曲。标准差和四分位距对于大数据集而言更加实用。
For a sample of n values, the standard deviation s is:
s = √( (Σx² − n·x̄²) / (n − 1) )
In A-Level statistics, the formula for sample variance uses (n − 1) in the denominator, not n. Your calculator's 'sx' or 's' button gives this sample standard deviation directly.
在 A-Level 统计中，样本方差公式分母使用 (n − 1)，而不是 n。计算器上的 “sx” 或 “s” 按键直接给出样本标准差。
The interquartile range (IQR) is defined as:
IQR = Q₃ − Q₁
The IQR measures the middle 50% of the data and is not sensitive to outliers. If comparing two locations, the one with the smaller standard deviation or IQR has more consistent weather.
四分位距衡量中间 50% 数据的分布范围，不受异常值影响。如果比较两个地点，标准差或四分位距更小的地点，其天气变化更稳定。
6. Quartiles, Percentiles and Box Plots | 四分位数、百分位数与箱线图
Quartiles divide data into four equal parts. Q₁ is the 25th percentile, Q₂ is the median (50th percentile), and Q₃ is the 75th percentile. Percentiles are often used to show the relative standing of one observation within the whole data set.
四分位数将数据分成四个相等部分。Q₁ 是第 25 百分位数，Q₂ 是中位数（第 50 百分位数），Q₃ 是第 75 百分位数。百分位数常用于表明某个观测值在整个数据集中的相对位置。
To find quartiles from a list of data, first arrange the data in ascending order. For the lower quartile, look at the median of the lower half of the data; for the upper quartile, look at the median of the upper half. Different exam boards have slightly different conventions, so always follow the method taught in your specification.
要从数据列表中求四分位数，首先按升序排列数据。下四分位数取下半部分数据的中位数，上四分位数取上半部分数据的中位数。不同考试局的约定略有差异，请始终遵循你所在考纲讲授的方法。
A box plot displays the minimum, Q₁, median, Q₃ and maximum values. If you are given a box plot, you can read all five values directly. Box plots are especially useful for comparing two large data sets side by side.
箱线图展示最小值、Q₁、中位数、Q₃ 和最大值。如果给出箱线图，你可以直接读出这五个数值。箱线图特别适合并排比较两个大数据集。
7. Outliers and Skewness | 异常值与偏态
An outlier is a data value that is unusually large or unusually small. In A-Level statistics, a common definition uses the IQR:
异常值是指异常偏大或偏小的数据值。在 A-Level 统计中，一个常用定义基于四分位距：
Outlier: x < Q₁ − 1.5 × IQR or x > Q₃ + 1.5 × IQR
If a question gives you raw data, you should first calculate Q₁, Q₃ and IQR, then test each potential extreme value using this rule. Box plots often show outliers as separate crosses or dots beyond the whiskers.
如果题目给出原始数据，你应先计算 Q₁、Q₃ 和 IQR，然后用此规则检验每个可能的极值。箱线图通常将异常值以独立的叉号或圆点画在须线之外。
Skewness ta are negatively skewed. When mean ≈ median, the distribution is approximately symmetric.
偏度衡量分布的对称性。若中位数小于平均数，数据呈正偏态；若中位数大于平均数，数据呈负偏态；若平均数约等于中位数，分布近似对称。
You can also assess skewness using quartiles:
If Q₃ − Q₂ > Q₂ − Q₁, the data are positively skewed.
If Q₃ − Q₂ < Q₂ − Q₁, the data are negatively skewed.
8. Correlation and Linear Regression | 相关性与线性回归
Scatter graphs can reveal relationships between two variables, such as temperature and sunshine hours. The product-moment correlation coefficient (PMCC), r, measures the strength and direction of a linear relationship.
散点图可以揭示两个变量之间的关系，例如气温与日照时数之间的关系。积差相关系数（PMCC），记作 r，衡量线性关系的强弱与方向。
−1 ≤ r ≤ 1
If r is close to 1, there is strong positive correlation; if r is close to −1, strong negative correlation. If r is close to 0, there is little or no linear correlation. Remember: correlation does not imply causation. A high r may be due to a lurking variable, such as season.
若 r 接近 1，则存在强正相关；若 r 接近 −1，则存在强负相关；若 r 接近 0，则几乎没有线性相关。请记住：相关不等于因果。高 r 值可能是由潜伏变量（如季节）造成的。
The equation of the regression line of y on x is written as:
y = a + bx
Here, b is the gradient and a is the intercept. Use your calculator's linear regression mode to find a and b from raw data, then use the equation to estimate y for a given x. But be careful: extrapolation, or estimating outside the data range, is unreliable.
其中 b 是斜率，a 是截距。使用计算器的线性回归模式，从原始数据求出 a 和 b，然后利用方程估计给定 x 对应的 y。但请注意：外推法，即估计超出数据范围的值，是不可靠的。
9. Time Series and Trends | 时间序列与趋势
Weather data are recorded over time, so time series plots are very common in large data set questions. A time series graph plots the variable against time in chronological order.
天气数据是随时间记录的，因此时间序列图在大数据集题目中非常常见。时间序列图将变量按时间顺序绘制在坐标轴上。
In a time series, you may be asked to identify: A seasonal variation is a pattern that repeats at regular intervals, such as colder temperatures every winter. A trend is the underlying long-term movement, often found using a moving average.
在时间序列中，你可能需要识别：季节性变化是定期重复的模式，例如每年冬季气温降低；趋势是潜在的长期运动方向，通常用移动平均法找出来。
To calculate a moving average for a set of n values with period m, take the mean of each block of m consecutive values. For example, the 4-point moving average uses blocks (x₁, x₂, x₃, x₄), then (x₂, x₃, x₄, x₅), and so on.
要计算周期为 m 的移动平均值，取每 m 个连续值的平均数。例如，4 点移动平均先取 (x₁, x₂, x₃, x₄) 的平均值，再取 (x₂, x₃, x₄, x₅) 的平均值，以此类推。
If seasonal variation is present, adding a moving average trend line to the graph helps separate the trend from the seasonal effect. You should also be able to comment on whether the trend is increasing, decreasing or stable.
如果存在季节性变化，在图上添加移动平均趋势线有助于将趋势与季节效应分开。你还应能够说明趋势是在上升、下降还是保持平稳。
10. Coding | 数据编码
Coding is a technique used to simplify calculations by subtracting a constant or dividing by a constant. In the large data set, weather variables may be coded, for example by converting temperatures in tenths of a degree or by subtracting the baseline value.
编码是一种通过减去常数或除以常数来简化计算的技术。在大数据集中，天气变量可能经过编码，例如将温度转换为十分之一度，或减去一个基准值。
If a variable x is coded to y using a linear transformation, for example:
y = (x − a) / b
then the mean and standard deviation of y can be related back to the mean and standard deviation of x. In general:
若变量 x 通过线性变换编码为 y，例如：
y = (x − a) / b
那么 y 的平均数和标准差可以还原为 x 的平均数和标准差。一般规则为：
x̄ = a + bȳ
sₓ = b·sᵧ, if b > 0
Notice that subtracting constant a shifts both the mean and median, but does not affect the standard deviation or IQR. Dividing by constant b scales the spread. If you only need the correlation coefficient or standardised measures, coding does not change r.
注意：减去常数 a 会平移平均数和中位数，但不影响标准差或四分位距；除以常数 b 会缩放数据的离散程度。若只需相关系数或标准化量，编码不会改变 r 的值。
11. Comparing Data Sets | 数据集比较
Exam questions often ask you to compare two large data sets, such as rainfall in London and rainfall in Beijing. A full comparison should include a measure of location and a measure of spread, with both a numerical value and a contextual interpretation.
考试题目经常要求你比较两个大数据集，例如伦敦和北京的降雨量。完整的比较应包含一个集中趋势量和一个离散程度量，并且要同时写出数值和情境解释。
You can compare: With two samples, even if the means are equal, the standard deviations may be different. A smaller standard deviation indicates that the data are more tightly clustered around the mean, implying more consistent weather.
你可以比较：两个样本，即使平均数相同，标准差也可能不同。标准差较小意味着数据更紧密地聚集在平均数周围，说明天气更稳定。
- Mean and median: compare the typical values.
- 平均数与中位数：比较典型取值水平。
- IQR or standard deviation: compare the variability.
- 四分位距或标准差：比较波动程度。
- Overall range: compare the extremes, but note the effect of outliers.
- 全距：比较极端值范围，但要注意异常值的影响。
Always write your comparison in context. Instead of saying 'Location A has a larger mean', say 'the mean daily temperature in Perth is higher than in Lerwick, suggesting a warmer climate overall'.
写比较时一定要结合情境。不要说“地点 A 的平均数更大”，而要说“珀斯的每日平均气温高于勒威克，表明其整体气候更温暖”。
12. Exam-Style Command Words | 考试常用指令词
To maximise marks, you need to respond to the command word accurately. Large data set questions often use the following terms:
要获得高分，你需要准确回应题目中的指令词。大数据集题目常见以下术语：
| Command Word | What You Must Do |
| State / Write down | Give the answer with no working needed. |
| Calculate | Show relevant working; use the calculator correctly. |
| Interpret | Explain the statistical meaning in context. |
| Compare | Give similarities or differences using measures and context. |
| Describe the trend | Refer to direction, shape and possible outliers. |
| Explain why | Give a mathematical or contextual reason. |
Descriptive command verbs require context. A bare statistic often loses the final mark.
Before submitting, check your calculator model, remember to write the unit (°C, mm, hours, knots) in your answer, and round to a sensible degree of accuracy, usually 3 significant figures or 1 decimal place as directed by the question.
交卷前，请检查计算器型号是否正确，答案中要写上单位（°C、mm、小时、节），并按题目要求保留合理的精度，通常保留 3 位有效数字或 1 位小数。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图