Government Operational Research Service (GORS) Assessment Data Analysis, Modelling and Analytical Tools Syllabus

Data Analysis, Modelling and Analytical Tools syllabus — full chapter and topic list

Expand any chapter to see its topics and sub-topics. This is the whole examinable outline for Data Analysis, Modelling and Analytical Tools in Government Operational Research Service (GORS) Assessment, not a summary of it.

1. 1

   Working with data

   4 topics

   • Data sources across government
     • Administrative, survey and operational data
     • Open data and secondary sources
   • Data quality, cleaning and preparation
   • Exploratory data analysis
     • Summary statistics and distributions
     • Outliers and missing data
   • Data ethics, privacy and protective characteristics
2. 2

   Statistical and machine learning modelling

   4 topics

   • Supervised learning fundamentals
     • Classification and regression
     • Training, validation and test splits
   • Unsupervised learning and clustering
   • Model evaluation and overfitting
   • Interpretability and responsible AI in government
3. 3

   Analytical tools and software (RAP)

   4 topics

   • Reproducible Analytical Pipelines (RAP)
     • Principles and benefits
     • Version control with Git
   • Programming for analysis in R and Python
   • Spreadsheets and modelling in Excel
   • Specialist OR and simulation software
4. 4

   Communicating data and results

   4 topics

   • Data visualisation principles and good practice
     • Choosing the right chart
     • Accessibility and clarity
   • Dashboards and interactive reporting
   • Communicating uncertainty visually and verbally
   • Storytelling with data for decision-makers

Data Analysis, Modelling and Analytical Tools flashcards for Government Operational Research Service (GORS) Assessment

19 of 64 cards from the Data Analysis, Modelling and Analytical Tools deck — real questions with worked answers.

1. Name four major categories of data sources used across UK government analysis.

   Administrative data (collected for operational/service delivery, e.g. tax, benefits records), survey data (e.g. ONS Labour Force Survey), census data (decennial population count), and open/published data (e.g. data.gov.uk). Increasingly also commercial, sensor/transactional, and web-scraped sources.

2. What is the difference between administrative data and survey data?

   Administrative data is a by-product of running a service or process (not collected for research), giving large coverage but potentially incomplete/biased fields. Survey data is purpose-collected from a sample using designed questions, allowing controlled measurement but at higher cost and with sampling error.

3. In data analysis, distinguish primary data from secondary data.

   Primary data is collected first-hand by the analyst for the specific question. Secondary data was collected by someone else for another purpose and is reused. Secondary data is cheaper and faster but may not exactly fit the need.

4. What are the commonly cited dimensions of data quality?

   Accuracy, completeness, consistency, timeliness, validity, and uniqueness (no duplication). Some frameworks add relevance, accessibility, and coherence/comparability (the latter emphasised in official statistics).

5. List the typical steps in a data cleaning and preparation process.

   1) Profile/inspect the data; 2) handle missing values; 3) correct invalid/out-of-range values and fix data types; 4) remove or resolve duplicates; 5) standardise formats and units; 6) detect and treat outliers; 7) validate against business rules; 8) document the transformations.

6. Compare the three common strategies for handling missing data.

   Deletion (listwise/pairwise) removes records but loses information and can bias results if data are not Missing Completely At Random. Imputation fills gaps using mean/median/mode, regression, or model-based (e.g. multiple imputation) methods. Flagging/indicator adds a missingness marker so models can use the pattern itself.

7. Define MCAR, MAR and MNAR missingness mechanisms.

   MCAR (Missing Completely At Random): missingness unrelated to any data. MAR (Missing At Random): missingness depends only on observed variables. MNAR (Missing Not At Random): missingness depends on the unobserved value itself (hardest to handle, can bias estimates).

8. What is Exploratory Data Analysis (EDA) and what is its purpose?

   EDA is the initial, mostly visual and summary-statistic investigation of data to understand its structure, spot patterns, anomalies and outliers, check assumptions, and form hypotheses before formal modelling. Associated with John Tukey.

9. State the formulas for the sample mean and sample variance.

   Mean: $\bar{x} = \frac{1}{n}\sum_{i=1}^{n} x_i$. Sample variance: $s^{2} = \frac{1}{n-1}\sum_{i=1}^{n}(x_i - \bar{x})^{2}$, with sample standard deviation $s = \sqrt{s^{2}}$.

10. How is the interquartile range (IQR) defined and how is it used to flag outliers?

    $\text{IQR} = Q_3 - Q_1$. A common rule flags values below $Q_1 - 1.5\times\text{IQR}$ or above $Q_3 + 1.5\times\text{IQR}$ as outliers. This is the basis of the boxplot whisker rule.

11. Give the formula for Pearson's correlation coefficient and its range.

    $r = \frac{\sum (x_i - \bar{x})(y_i - \bar{y})}{\sqrt{\sum (x_i - \bar{x})^{2}}\,\sqrt{\sum (y_i - \bar{y})^{2}}}$, where $-1 \leq r \leq 1$. It measures the strength and direction of a linear relationship only.

12. What is the difference between correlation and causation?

    Correlation means two variables move together statistically; causation means one variable directly produces a change in the other. Correlation can arise from coincidence, reverse causation, or a confounding third variable, so it does not imply causation.

13. In UK data ethics, what are the nine protected characteristics under the Equality Act 2010?

    Age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex, and sexual orientation.

14. Name the six Data Protection Principles under UK GDPR.

    1) Lawfulness, fairness and transparency; 2) purpose limitation; 3) data minimisation; 4) accuracy; 5) storage limitation; 6) integrity and confidentiality (security). A seventh overarching principle is accountability.

15. What guidance underpins ethical data use for UK government analysts?

    The Data Ethics Framework (CDDO/DSIT) with its principles of transparency, accountability and fairness, alongside the ONS/UK Statistics Authority approach. The Aqua Book governs quality assurance of analysis, and the Code of Practice for Statistics sets Trustworthiness, Quality and Value (TQV).

16. Distinguish anonymisation from pseudonymisation.

    Anonymisation irreversibly removes the ability to identify individuals, so the data falls outside data-protection law. Pseudonymisation replaces identifiers with tokens/keys; re-identification is still possible with the key, so it remains personal data under UK GDPR.

17. Define supervised learning and give its two main task types.

    Supervised learning trains a model on labelled input-output pairs to predict a target. The two main tasks are classification (predicting a discrete category) and regression (predicting a continuous numeric value).

18. Write the simple linear regression model and what each term represents.

    $y = \beta_0 + \beta_1 x + \varepsilon$, where $\beta_0$ is the intercept, $\beta_1$ the slope (change in $y$ per unit $x$), and $\varepsilon$ the random error term. Coefficients are typically fitted by ordinary least squares.

19. What does Ordinary Least Squares (OLS) minimise?

    It minimises the residual sum of squares: $\text{RSS} = \sum_{i=1}^{n}(y_i - \hat{y}_i)^{2}$, the sum of squared differences between observed and predicted values.

See more Data Analysis, Modelling and Analytical Tools flashcards →

Planning Data Analysis, Modelling and Analytical Tools for Government Operational Research Service (GORS) Assessment

Data Analysis, Modelling and Analytical Tools (Government Operational Research Service (GORS) Assessment) FAQ