Difficulty: Beginner | Prerequisites: Basic understanding of what data is and how it is collected. No programming knowledge required.
This topic sits at the intersection of data analysis, computing and society. It covers three ways that data collection and usage have expanded beyond traditional methods: big data (processing massive datasets with parallel systems), crowdsourced data (ordinary people contributing to scientific research), and open data (freely available datasets anyone can use). You should already be comfortable with the general idea of data collection and the basics of how computers process information. If you have covered earlier units on data representation and the internet, you are in good shape.
Modern computing lets us collect and process data at scales that were not possible before. "Big data" requires scalable, parallel systems to handle the volume. "Crowdsourced data" relies on everyday people (citizen scientists) to gather observations. "Open data" means datasets that anyone can freely access, use and build on. Together, these three ideas are reshaping how research, government and industry solve problems.
Citizen Science
The growing practice of public involvement in the gathering, analysing or sorting of scientific data for research purposes.
In simple terms, this means ordinary people help scientists collect data, often through apps or volunteer programmes. Think of birdwatchers logging sightings that researchers then use to track migration patterns.
Crowdsourcing
A method of obtaining services, ideas or data from a large, open network of people, usually via the internet.
In simple terms, instead of hiring a small team, you ask thousands of people online to each contribute a small piece. Polls, reviews and Wikipedia edits are all forms of crowdsourcing.
Open Access
Free access to information and unrestricted use of electronic resources for everyone.
In simple terms, content that anyone can read, download or use without paying or asking permission. Academic journals that publish without a paywall are open access.
Open Data
Data that can be freely used, shared and built upon by anyone, anywhere, for any purpose.
In simple terms, think of government datasets (census figures, transport schedules, weather records) published online for anyone to download and analyse. The key difference from open access is that open data specifically refers to raw, machine-readable datasets rather than finished publications.
Parallel Algorithms
An algorithm designed to perform multiple operations at the same time, dividing work across processors.
In simple terms, instead of one worker doing every task in sequence, you split the job among many workers who all work simultaneously. This is how big data systems crunch enormous datasets in reasonable time.
Scalability
The measure of a system's ability to increase or decrease in performance and cost in response to changes in processing demands.
In simple terms, a scalable system handles growth well. If a website can serve ten users or ten million users without falling over, it scales. Scalability matters for big data because the volume of data keeps growing.
What makes data "big"
Volume so large that traditional tools (a single spreadsheet, a single computer) cannot process it in a useful timeframe
Generated continuously from sensors, social media, transactions, scientific instruments and more
Requires specialised infrastructure to store, move and analyse
Parallel systems and parallel algorithms
Rather than processing data sequentially on one machine, the workload is split across many processors or machines running simultaneously
A parallel algorithm divides a problem into smaller sub-problems, solves them at the same time, then combines the results
This is the core technique that makes big data analysis feasible
Scalability in big data
A system is scalable if it can handle increasing data volumes by adding more resources (processors, storage, bandwidth) without a redesign
Good scalability means performance stays acceptable as demands grow
Poor scalability means the system slows down or fails under heavier loads
Big data in medicine
Hospitals and research institutions generate enormous datasets from patient records, imaging, genomics and clinical trials
Parallel processing allows researchers to find patterns across millions of records, for example identifying risk factors for diseases or predicting treatment outcomes
What crowdsourcing looks like in practice
An open call to a large group of people (usually online) to contribute data, ideas or labour
Contributors are typically volunteers, not paid specialists
Examples: online polls, product reviews, collaborative wikis, tagging images for machine learning
Citizen science as a form of crowdsourcing
Volunteers gather empirical observations or use equipment (cameras, apps, measuring tools) to create scientific data
Researchers design the project, set protocols and analyse the results; the public provides the scale
Example: Pokémon Go-style apps that encourage players to photograph and log wildlife sightings, contributing to biodiversity databases
The data analysis process in citizen science
Collect or choose data: volunteers record observations or capture images in the field
Clean and filter data: researchers compare submissions for accuracy; studies suggest 51 to 62 per cent of citizen science comparisons yield accurate data, depending on the method
Visualise and find patterns: aggregated observations reveal trends (for example, a species appearing in new regions)
New information: the collective dataset grows beyond what any small research team could gather alone
Accuracy considerations
Citizen science data is not automatically reliable; verification steps and cross-checking are essential
Large sample sizes help compensate for individual errors
Core principle
Data that is published in a format anyone can access, use, modify and redistribute without restrictions
Typically machine-readable (CSV, JSON, API endpoints), not just a PDF you can view
Government agencies, international organisations and research institutions are the largest publishers of open data
Open data vs. open access
Open access usually refers to published works (journal articles, reports) being free to read
Open data refers to the raw, underlying datasets being free to download and reuse
Both remove barriers, but open data emphasises reusability and machine-readability
Why open data matters
Transparency: the public can inspect how decisions were made
Innovation: developers and researchers build new tools and analyses on top of published datasets
Accountability: open crime statistics, spending records or environmental monitoring data let citizens hold institutions to account
Examples
Government budget data published for public scrutiny
Transport timetables released so developers can build journey-planning apps
Wikipedia, which allows anyone to access and edit its content (though Wikipedia is more accurately described as an open-access, crowdsourced knowledge base rather than a pure open-data repository)
All three topics (big data, crowdsourced data and open data) follow the same four-step data analysis process. Understanding this cycle is essential for the exam.
Step 1: Collect or choose data
Decide what data you need and where it comes from
In citizen science, volunteers record observations; in big data, sensors and systems generate it automatically; in open data, you select from published datasets
Step 2: Clean and filter data
Remove errors, duplicates and irrelevant entries
Verify accuracy (especially important with crowdsourced data, where contributors may make mistakes)
Decide which subset of the data is relevant to the question being investigated
Step 3: Visualise and find patterns
Use charts, graphs, maps or statistical tools to spot trends, clusters or outliers
Patterns may reveal correlations (for example, a link between location and species diversity)
Step 4: Draw new information and conclusions
Interpret the patterns to answer the original question
Share findings, contribute back to the dataset, or use the conclusions to inform decisions
Big data in healthcare: Hospitals use parallel processing to analyse millions of patient records, finding patterns that help predict disease outbreaks or personalise treatments.
Citizen science in ecology: Projects like eBird and iNaturalist rely on volunteer birdwatchers and naturalists to log sightings, building global biodiversity maps that no single research team could create.
Open data in civic life: City governments publish transport and crime data so that independent developers can build apps, journalists can investigate trends, and residents can make informed decisions about where to live.
Students often think "big data" just means "a lot of data." It does not. Big data specifically implies datasets so large or complex that they require parallel processing and scalable infrastructure, not simply a large spreadsheet.
Students sometimes confuse open data with open access. Open access is about free access to finished publications. Open data is about raw, reusable datasets. They overlap, but they are not the same thing.
Students sometimes assume crowdsourced data is unreliable and therefore useless. In practice, large sample sizes and verification protocols can produce accurate results, though the data does require more cleaning than professionally collected data.
Students occasionally treat citizen science and crowdsourcing as identical. Citizen science is a specific application of crowdsourcing focused on scientific research. Crowdsourcing is the broader concept and includes non-scientific activities like product reviews or collaborative writing.
⚠️ Be able to define and distinguish between big data, crowdsourced data and open data. Expect a question that gives you a scenario and asks which category it falls into.
⚠️ Know the four steps of the data analysis process (collect, clean, visualise, conclude) and be ready to apply them to any of the three topics.
⚠️ Understand why parallel algorithms are necessary for big data. The exam may ask what happens when data volume exceeds a single system's capacity.
⚠️ Be prepared to explain the relationship between citizen science and crowdsourcing (citizen science is a subset).
⚠️ Know the difference between open data and open access. This distinction is a common exam trap.
True or False: Open data and open access mean exactly the same thing.
False. Open access refers to free access to published works; open data refers to raw, reusable datasets.
Fill in the blank: A __________ algorithm divides a task across multiple processors so they can work at the same time.
Parallel.
True or False: Citizen science data is always accurate because many people contribute.
False. Large sample sizes help, but verification and cleaning are still necessary.
Fill in the blank: __________ is the measure of a system's ability to handle increasing demands by adding resources.
Scalability.
True or False: Crowdsourcing is limited to scientific research.
False. Crowdsourcing is a broad concept. Citizen science is the scientific subset.
Q: A weather service collects temperature readings from millions of sensors every second. What type of data challenge does this represent, and what computing approach makes it feasible to process?
A: This is a big data challenge. The volume and speed of incoming data require parallel algorithms running on scalable infrastructure to process in a useful timeframe.
Q: A university asks volunteers across the country to photograph and log wildflower sightings using a mobile app. What type of data collection is this, and what is one concern about the quality of the data?
A: This is citizen science (a form of crowdsourcing). A key concern is data accuracy, since volunteers may misidentify species or record locations incorrectly. Researchers must clean and verify submissions.
Q: A city government publishes its annual budget as a downloadable CSV file that anyone can analyse. Is this an example of open data, open access, or both? Explain.
A: This is open data. The raw dataset is machine-readable and freely available for anyone to use and build on. It would also qualify as open access in the broader sense that it is free to view, but the defining feature here is the reusable, raw dataset.
Q: Explain why a single powerful computer might not be enough to analyse a big dataset, even if that computer has a fast processor.
A: A single processor handles tasks sequentially. When the dataset is large enough, sequential processing takes too long to be practical. Parallel algorithms distribute the work across many processors simultaneously, reducing the total time required. Scalability also matters: a single machine cannot add more processors on the fly, whereas a distributed system can.
Q: Name two benefits of open data for society.
A: Transparency (the public can see how decisions are made and hold institutions accountable) and innovation (developers and researchers can build new tools, analyses and applications on top of freely available datasets).
This material connects directly to earlier units on data representation and the internet. The internet is the infrastructure that makes crowdsourcing and open data distribution possible at scale. Understanding how data is stored and transmitted (binary, compression, protocols) gives context for why scalability and parallel processing matter when datasets grow.
It also links to privacy and security topics covered later in most CSP courses. Open data raises questions about what should be public and what should remain private, and big data analysis can inadvertently reveal personal information even from anonymised datasets.
big data, open data, open access, crowdsourcing, crowdsourced data, citizen science, parallel algorithms, parallel processing, parallel systems, scalability, scalable systems, data analysis process, data collection, data cleaning, data visualisation, distributed computing, volunteer science, public datasets, government data, machine-readable data, CSP Unit 9, Computer Science Principles, AP CSP data