Introduction to the Large Data Set (LDS)

Welcome to one of the most unique parts of your A-level Statistics course! Unlike other chapters where you learn formulas to solve abstract problems, the Large Data Set is all about becoming a "data detective." You are expected to become familiar with a specific set of real-world data before you even walk into the exam room.

In the AQA 7357 specification, the examiners want to see that you can handle real data—which is often messy, missing pieces, or full of surprises—rather than just the "perfect" numbers usually found in textbooks.

What is the Current Data Set?

For exams from 2020 onwards, the prescribed data set focuses on Car Data. This is a collection of records for various vehicles, and you should be comfortable navigating its different "fields" (the columns in a spreadsheet). Quick tip: You don't need to memorize every single number, but you do need to know what the variables represent!

The Key Variables (Data Fields)

According to the official syllabus, the "Car data" sheet contains several important categories:

  • ReferenceNumber: A unique ID for each record.
  • Make: There are only five specific makes included in the set.
  • PropulsionTypeId: How the car is powered (e.g., Petrol, Diesel, Electric).
  • BodyTypeId: The shape/style of the car.
  • GovRegion: The English region where the car's keeper lives (based on postcodes).
  • KeeperTitleId: This often indicates the gender of the keeper (based on DVLA titles like Mr, Mrs, Ms).
  • YearRegistered: When the car first hit the road.
  • EngineSize: Measured in \( cm^3 \).
  • Mass: Measured in \( kg \). Crucial Note: This value includes a standard \( 75 \text{ kg} \) for the driver!
  • Emissions: This includes \( CO_2 \), \( CO \), \( NO_X \), particulates, and hydrocarbons.

Data Cleaning and Missing Values

In a perfect world, every car would have every bit of data recorded. In the real world, data is often missing. This is a "live feature" of the AQA data set that you are expected to understand.

Cleaning the data is the process of dealing with these gaps or errors before you start your analysis. For example:

  • Particulates: This data is usually only recorded for diesel vehicles. If you look at a petrol car in the spreadsheet, this cell might be empty or show an error.
  • Why does this matter? If an exam question asks why a certain average might be misleading, you could point out that missing data for certain engine types might skew the results.
Quick Review: What is "Cleaning"?

Data cleaning involves:
1. Identifying missing values (often marked as \( N/A \)).
2. Spotting outliers (values that look impossibly high or low).
3. Deciding whether to remove those records or ignore the empty fields.

How You Will Use This in the Exam

You won't be given the entire spreadsheet in the exam. Instead, you might be given a subset (a small part) of the data or a summary table/graph based on it. The examiners expect you to:

  • Interpret graphs: Look at scatter diagrams or histograms of the car data and describe what they show about correlation or distribution.
  • Use your calculator: You should be able to calculate summary statistics (like the mean \( \bar{x} \) or standard deviation \( \sigma \)) for a small sample of the data provided in the question.
  • Show "Familiarity": Questions might ask you to comment on a specific feature of the data set that you should already know, such as which regions are included or how mass is calculated.

Don't worry if this seems tricky at first! You don't need to be a car expert. You just need to spend some time exploring the spreadsheet using software like Excel or Google Sheets. Try filtering the data—for example, look only at "Ford" cars or only at cars registered in "London"—to see how the averages change.

Common Pitfalls to Avoid

1. Forgetting the Driver: Remember that the Mass field includes \( 75 \text{ kg} \) for the driver. If a question asks for the mass of the vehicle alone, you must subtract \( 75 \).
\( \text{Vehicle Mass} = \text{Data Set Mass} - 75 \)

2. Generalizing too much: Correlation does not imply causation. Just because older cars might have higher emissions in the data set doesn't mean every old car is "dirtier" than a new one; it's just a trend in this specific data.

3. Ignoring Units: Always check your units! Engine size is in \( cm^3 \), mass is in \( kg \), and different emissions might be measured in different units (like \( g/km \) or \( g/kWh \)).

Key Takeaways

  • The Large Data Set for AQA focuses on Car Data (from 2020 onwards).
  • It contains categorical data (Make, Region, Title) and quantitative data (Mass, Engine Size, Emissions).
  • Mass includes a \( 75 \text{ kg} \) driver.
  • Missing data is a key theme; you must understand that not every field is filled for every car (e.g., particulates for diesel only).
  • You should use technology (spreadsheets and your statistical calculator) to practice analyzing subsets of this data before the exam.

Note: For more information on how to display this data graphically or how to use it in hypothesis tests, see the chapters on Data presentation and interpretation and Statistical hypothesis testing.