Course outline
Module 1
Module 1: What Open Data Is and Why It Matters
What Open Data Is and Why It Matters
Knowledge check
Q1. Which of the following best describes open data?
- A. Data that is stored in a password-protected database accessible only to employees
- B. Data that anyone can freely access, use, modify, and share, subject at most to requirements to credit the source or keep it open ✓
- C. Data that is sold to private companies at a discounted rate
- D. Data collected only by government agencies for internal planning purposes
Why: Open data is defined by being freely available for anyone to use, reuse, and redistribute, typically under an open license, with at most minimal conditions such as attribution.
Q2. Which of the following is a key characteristic that makes data truly 'open'?
- A. It is available only after paying a subscription fee
- B. It is published in a machine-readable format under an open license ✓
- C. It can only be viewed as a PDF document on a website
- D. It is restricted to users who register with a government email address
Why: Open data should be machine-readable (such as CSV or JSON) so it can be processed automatically, and it should carry an open license that permits reuse.
Q3. Why does open data matter for society?
- A. It allows governments to keep citizens uninformed about public spending
- B. It promotes transparency, accountability, innovation, and informed decision-making ✓
- C. It eliminates the need for private research organizations entirely
- D. It guarantees that all data will be accurate and error-free
Why: Open data helps citizens hold institutions accountable, enables businesses and researchers to build new tools and insights, and supports better-informed public decisions.
Q4. Which example is most likely to be considered open data?
- A. A company's confidential internal sales records
- B. A city's published list of public transit stop locations, released under an open license ✓
- C. A hospital's private patient medical histories
- D. A research team's unpublished experimental results kept in a private lab notebook
Why: Public transit stop locations released openly for reuse fit the definition of open data, while confidential business, medical, and unpublished research data do not.
Q5. What is a common reason organizations hesitate to publish open data?
- A. Open data is always illegal in every country
- B. Concerns about privacy, security, and the cost of preparing data for publication ✓
- C. Open data cannot be stored on computers
- D. Open data always reduces the quality of the original dataset
Why: Organizations often worry that releasing data could expose personal information or sensitive details, and preparing data responsibly for release takes time and resources.
Module 2
Module 2: Open Data Licences and Legal Basics
Open Data Licences and Legal Basics
Knowledge check
Q1. What is the main purpose of an open data licence?
- A. To restrict who can view a dataset
- B. To specify the legal terms under which others may use, reuse, and share the data ✓
- C. To encrypt a dataset so it can only be read by its owner
- D. To register a dataset as a trademark
Why: An open data licence grants legal permission to use, reuse, and redistribute data, and sets out any conditions attached to that use.
Q2. Which licence is most commonly recommended for open government data in the UK?
- A. Creative Commons Attribution 4.0 (CC BY 4.0)
- B. Open Government Licence (OGL) v3.0 ✓
- C. GNU General Public Licence (GPL)
- D. All rights reserved copyright
Why: The Open Government Licence v3.0 is the standard licence for UK public sector information, allowing free use and reuse with attribution.
Q3. What does the 'BY' (Attribution) condition in a Creative Commons licence require?
- A. Users must pay a fee before reusing the data
- B. Users must credit the original creator when reusing the data ✓
- C. Users may only use the data for non-commercial purposes
- D. Users must share any derivative work under the same licence
Why: The BY condition requires that the original author or source be credited whenever the work is used or redistributed.
Q4. Why is it important to check a dataset's licence before reusing it?
- A. Licences determine the file format of the data
- B. Using data without the correct permissions can lead to legal risk, even if the data is publicly accessible ✓
- C. Licences automatically update the data each year
- D. Only licensed data can be opened in spreadsheet software
Why: Public availability does not equal permission to reuse; the licence defines what you are legally allowed to do, and ignoring it can create legal exposure.
Q5. Which statement best describes 'public domain' data?
- A. Data that is kept secret by government agencies
- B. Data that has no copyright or other restrictions and can be used freely by anyone ✓
- C. Data that can only be used by public universities
- D. Data that must be purchased before use
Why: Public domain material is not restricted by copyright or similar rights, so it can be used freely without needing a licence.
Module 3
Module 3: Finding Open Data Sources
Finding Open Data Sources
Knowledge check
Q1. Which of the following best describes open data?
- A. Data that is only available to paying subscribers of a private company
- B. Data that anyone can freely access, use, modify, and share, usually with minimal restrictions ✓
- C. Data that is stored on a public cloud server regardless of its license
- D. Data that has been anonymized so that it cannot be traced to individuals
Why: Open data is defined by its free availability and permissive reuse terms, typically under an open license, allowing anyone to access, use, modify, and share it.
Q2. Which type of organization is a common publisher of open government data?
- A. Private advertising agencies
- B. National, state, or local government agencies ✓
- C. Social media platforms
- D. Individual online shopping stores
Why: Government agencies at national, regional, and local levels routinely publish open datasets on topics such as census statistics, transportation, health, and budgets.
Q3. What is a data portal in the context of finding open data?
- A. A physical server room where data is stored
- B. A centralized website that catalogs and provides access to multiple open datasets ✓
- C. A password-protected database used only by internal staff
- D. A type of file format used for compressing large datasets
Why: Data portals, such as data.gov or city open data sites, act as centralized catalogs that organize and provide access to many datasets from one or more publishers.
Q4. Which file format is generally considered most open and machine-readable for tabular open data?
- A. A scanned PDF of a printed table
- B. A proprietary spreadsheet format that requires specific paid software
- C. A CSV (comma-separated values) file ✓
- D. An image file showing a chart
Why: CSV is a plain-text, widely supported, machine-readable format that can be opened by almost any software, making it a preferred format for open data.
Q5. When you find an open dataset, what should you check before reusing it?
- A. Only the color scheme of the publisher's website
- B. The license terms, metadata, and last-updated date ✓
- C. Whether the dataset has a catchy title
- D. The number of social media followers of the publisher
Why: Checking the license clarifies how you may use the data, metadata describes its content and source, and the update date tells you whether the data is current enough for your purpose.
Module 4
Module 4: Understanding Data Formats and Structure
Understanding Data Formats and Structure
Knowledge check
Q1. Which file format stores data as plain text with values separated by commas?
- A. XML
- B. CSV ✓
- C. PDF
- D. JPEG
Why: CSV (Comma-Separated Values) is a simple plain-text format where each line is a record and commas separate the individual fields, making it widely used for open datasets.
Q2. What does each row in a typical tabular dataset usually represent?
- A. A single column of metadata
- B. A single observation or record ✓
- C. The file's encoding type
- D. A chart or visualization
Why: In tabular data, each row is typically one observation or record (for example, one person, one transaction, or one weather reading), while each column holds a specific attribute of those records.
Q3. Which format uses nested tags to represent hierarchical data?
- A. CSV
- B. TXT
- C. XML ✓
- D. WAV
Why: XML (eXtensible Markup Language) uses opening and closing tags that can be nested, allowing it to represent hierarchical or structured relationships between data elements.
Q4. Why is JSON popular for sharing open data through web APIs?
- A. It can only store images
- B. It is lightweight and easy for both humans and machines to read, using key-value pairs ✓
- C. It cannot be parsed by programming languages
- D. It stores data only as binary code
Why: JSON (JavaScript Object Notation) is lightweight, human-readable, and maps naturally to data structures in most programming languages, which is why it is commonly used by web APIs.
Q5. What is a header row in a CSV file?
- A. A row containing the file's download date
- B. A row of column names that describes the data in each column ✓
- C. A row containing only empty values
- D. A row that marks the end of the file
Why: The header row is typically the first row of a CSV file and provides column names, so users know what each field represents without needing external documentation.
Module 5
Module 5: Cleaning and Preparing Open Datasets
Cleaning and Preparing Open Datasets
Knowledge check
Q1. Which of the following is the most common first step when cleaning an open dataset?
- A. Publishing the dataset online
- B. Inspecting the data to understand its structure, columns, and data types ✓
- C. Merging it with another dataset
- D. Creating visualizations of the final results
Why: Before making any changes, you need to understand what the data contains, how it is structured, and what each column represents. Inspection helps reveal problems such as wrong data types, unexpected values, or missing fields.
Q2. What is a missing value in a dataset?
- A. A row that has been duplicated
- B. A cell that contains no recorded value for a given field ✓
- C. A column that has been renamed
- D. A value that is outside the expected range
Why: A missing value means a field in a record has no data recorded. Missing values are common in open datasets and must be identified and handled, for example by removing the record, filling in a value, or noting the gap.
Q3. Why is it important to remove duplicate records from a dataset?
- A. Duplicates make files larger but have no other effect
- B. Duplicates can inflate counts and totals, leading to inaccurate analysis ✓
- C. Duplicates are required for data licensing
- D. Duplicates prevent the data from being opened in spreadsheet software
Why: When the same record appears more than once, totals, averages, and counts become distorted. Removing duplicates ensures that each real-world entity is counted only once in your analysis.
Q4. Which practice helps keep a record of the changes made while cleaning an open dataset?
- A. Overwriting the original file each time
- B. Deleting the original file after cleaning
- C. Keeping the raw data untouched and documenting each cleaning step ✓
- D. Sharing only the final version without any notes
Why: Keeping the original raw data unchanged and documenting each transformation makes your work reproducible and transparent. Others can verify your steps, and you can recover if a cleaning step turns out to be wrong.
Q5. A column called 'Age' contains the value 'unknown' for some rows and numbers for others. What is the best way to address this?
- A. Leave the text in place because it does not affect the data
- B. Delete the entire column from the dataset
- C. Convert 'unknown' to a proper missing value such as blank or NA so the column is consistently numeric ✓
- D. Replace every number in the column with the word 'unknown'
Why: Mixing text placeholders with numbers makes the column hard to analyze and can cause errors in calculations. Standardizing unknown entries to a true missing-value marker keeps the column numeric and clearly indicates where data is absent.
Module 6
Module 6: Analysing and Visualising Open Data
Analysing and Visualising Open Data
Knowledge check
Q1. Which of the following is the best first step when beginning to analyse an open dataset?
- A. Create a polished chart for publication
- B. Examine the dataset's structure, variables, and metadata to understand what it contains ✓
- C. Delete any rows that contain missing values
- D. Share the raw data on social media immediately
Why: Understanding the structure, variables, and metadata (including how the data was collected and what each field means) is essential before any analysis, so that conclusions are valid and errors are avoided.
Q2. What type of chart is most appropriate for showing how a value changes over time?
- A. Pie chart
- B. Line chart ✓
- C. Scatter plot with no time axis
- D. Histogram of a single categorical variable
Why: Line charts place time on the horizontal axis and show trends clearly, making them the standard choice for time-series data.
Q3. Why is it important to check the licence of an open dataset before visualising and publishing it?
- A. Licences determine the colour palette you must use
- B. Licences specify the conditions under which the data can be reused, attributed, or shared ✓
- C. Licences guarantee the data is accurate
- D. Licences are only needed for paid datasets
Why: Open data licences such as CC BY or the Open Government Licence set out how you may reuse and must credit the data, so checking them ensures lawful and proper use.
Q4. A dataset shows that ice cream sales and drowning incidents both rise in summer. What is the most accurate interpretation?
- A. Ice cream causes drowning incidents
- B. Drowning incidents cause ice cream sales to rise
- C. Both are correlated, likely due to a common factor such as hot weather, not a direct causal link ✓
- D. There is no relationship between the two variables
Why: Correlation does not imply causation. A confounding variable, here hot weather, can drive both trends, so analysts should avoid claiming a direct cause without further evidence.
Q5. What does it mean if a chart's y-axis does not start at zero?
- A. The chart is automatically incorrect
- B. The data contains negative values
- C. Visual differences may be exaggerated, so the reader should check the scale carefully ✓
- D. The data has not been cleaned
Why: A truncated y-axis can make small differences look large. It is not always wrong, but analysts should label the scale clearly and be transparent about it to avoid misleading viewers.
Module 7
Module 7: Ethics, Privacy and Responsible Reuse
Ethics, Privacy and Responsible Reuse
Knowledge check
Q1. Which of the following is the best description of open data?
- A. Data that is freely available to anyone to access, use, and share, usually under an open licence ✓
- B. Data that is kept private within a single organisation for internal use only
- C. Any spreadsheet stored on a public website, regardless of licence
- D. Data that has been encrypted so only authorised users can read it
Why: Open data is defined by being freely accessible and reusable by anyone, typically under an open licence that permits sharing and reuse. Private internal data and encrypted data are not open.
Q2. A dataset contains names, addresses, and dates of birth of patients. Before publishing it as open data, what is the most important step?
- A. Adding a watermark to each row of the dataset
- B. Converting the file to a different format such as PDF
- C. Removing or anonymising personal identifiers to protect individual privacy ✓
- D. Increasing the file size so it loads more slowly
Why: Personal identifiers can expose individuals to harm, so they must be removed or anonymised before release. Format changes and watermarks do not protect privacy.
Q3. Why can combining two anonymised open datasets sometimes lead to re-identification of individuals?
- A. Because anonymised data always becomes corrupted when merged
- B. Because overlapping attributes such as postcode, age, and gender can be linked to identify specific people ✓
- C. Because open licences forbid merging datasets
- D. Because merging datasets automatically deletes all records
Why: Quasi-identifiers like postcode, age, and gender can be matched across datasets, so individuals may be singled out even when direct identifiers are removed.
Q4. Which licence is generally considered suitable for open data because it permits reuse with attribution?
- A. All rights reserved, which forbids any copying
- B. Creative Commons Attribution (CC BY), which allows reuse if the creator is credited ✓
- C. A proprietary licence requiring a paid subscription
- D. No licence, since unlicensed material is always open
Why: CC BY allows others to use, share, and adapt the data as long as the source is credited, which fits open data principles. Restrictive or missing licences limit reuse.
Q5. When reusing open data responsibly, what should a researcher do?
- A. Use the data for any purpose without checking its source or limitations
- B. Check the source, licence, and known limitations, and cite the data appropriately ✓
- C. Assume the data is error-free because it is public
- D. Republish it under their own name without attribution
Why: Responsible reuse involves verifying provenance, respecting licence terms, understanding data quality limits, and giving proper credit to the original publisher.