presentsUNAIVERSE
The UN AI timeline

    Learn AI /AI Foundations /Data

    Data

    Detail

    In five seconds

    What it is

    Recorded observations, measurements or examples that people or systems can store, examine and use for a purpose.

    In plain words

    Data is recorded information that a person or computer can use.

    An everyday example

    Before a community meeting in Lima, Sofia asks volunteers to write one transport problem per slip of paper. Some slips say "bus late", some say "road flooded", and some include times and neighbourhood names. The team sorts the slips into groups, removes duplicates and corrects a few spelling mistakes so they can count the main issues. After that, they can make a summary for the meeting agenda.

    Where the story stops working

    Real data systems can handle millions of records automatically and may include images, audio or sensor readings, not just short written notes. Also, unlike slips of paper, digital data can be copied, linked and reused very easily, which creates governance, privacy and security issues the analogy does not show.

    Remember thisData is recorded material in a usable form. What a system can show or learn depends on what was recorded, how it was encoded and how it was prepared.

    Understand it

    Understand it

    Why it matters

    If you hear claims about AI, dashboards or evidence-based decisions, the first practical question is what data was used. Good decisions need relevant, lawful and well-managed data; poor or biased data can produce misleading outputs even when the software looks impressive.

    From observation to usable output

    Show that data starts as recorded observations, is prepared through several steps, and then becomes usable for analysis, reporting or AI tasks.

    1. 01

      Observe or record

      Information begins as something noticed or measured by a person or system.

      Draw source icons grouped in the first box: clipboard form, thermometer sensor and document page.

    2. 02

      Create data items

      Each observation becomes a stored item such as a row, file, message or image.

      Draw small cards in the second box with sample contents: '27°C', '12 May', 'bridge closed', and a simple image placeholder.

    3. 03

      Prepare the data

      Items are stored, organised and cleaned so they can be used reliably.

      Draw the third box with four small action chips labelled 'store', 'organise', 'clean', 'label'.

    4. 04

      Use the data

      Prepared data can support analysis, reporting or AI systems.

      Draw the final box with three outputs: bar chart, short report page and a tray labelled 'model input'.

    A left-to-right flow diagram with four boxes connected by arrows. Box 1 shows sources: a form icon, a sensor icon and a document icon. Box 2 shows recorded items as small cards containing examples such as a number, a date, a word and an image thumbnail. Box 3 shows preparation actions with labels: store, organise, clean, label. Box 4 shows outputs with icons for a chart, a report and an AI model input tray. Each arrow has a short label: collect, prepare, use.

    On hover or tap, each box reveals one example sentence: source example, data item example, preparation example and output example.

    Describe this without the diagram

    This diagram explains data as a process. First, information comes from sources such as forms, sensors or documents. Second, each observation is recorded as a data item, for example a temperature value, a date, a short text note or an image. Third, the items are prepared by storing them, organising them, cleaning errors and sometimes adding labels. Fourth, the prepared data is used to create outputs such as charts, reports or input for an AI model. The main message is that data is recorded information that becomes useful only after preparation for a specific purpose.

    With reduced motion: A static four-box flow chart shows the same sequence without animation: sources, recorded items, preparation actions and outputs, with labelled arrows reading collect, prepare and use.

    Data is the raw material many digital systems work with. It can be numbers, words, images, sound, locations, dates or labels. A single item might be a temperature reading, a line of text or a photo.

    What goes in depends on the task. A person, sensor, form or computer system records observations or events. Those records are then stored, organised, checked, combined or cleaned, which means fixing obvious errors and putting them into a usable form.

    What comes out can be many things. People may produce a chart, a list, a report or a decision with human review. An AI system may use data to find patterns, make a prediction or generate text, but the output depends strongly on the quality and relevance of the data it receives.

    Data is not the same as a dataset. Data can mean one item or many kinds of recorded information, while a dataset usually means a specific collection prepared for analysis or training. Data is also not the same as metadata, which is data about data, such as when a file was created or who entered a record.

    How “Notes on slips of paper before a meeting” maps onto the real thing

    In the storyIn the AI system
    each slip of paper with one problem written on ita data item or record
    times and neighbourhood names on the slipsdifferent features or fields in the data
    sorting, deduplicating and correcting the slipsorganising and cleaning data before use
    the summary for the agendaan output such as an analysis, report or model input

    A worked example

    A phone keyboard feature is built to suggest the next word in short messages.

    Starting from: Example training text includes lines such as: “see you tomorrow”, “see you soon”, “thank you”, “thank you very much” and “are you coming”.

    1. The text examples are encoded and stored as data.
    2. The system divides the text into smaller units, called tokens, and records which tokens tend to occur together.
    3. From the examples, it identifies that some words follow “thank you” more often than others.
    4. When a user types “thank you”, the system uses those learned patterns to rank possible next words.

    Result: The feature suggests the continuation that its data makes most likely. Different or newly collected data could produce a different ranking.

    In a United Nations workplace

    A communications officer at a field office with an unreliable connection is serving on a recruitment panel. Before the panel meets at close of business, she exports the applications from the approved recruitment system and records each candidate’s evidence against the rostered profile in a selection matrix.

    The exported fields, application text and entries in the selection matrix are all data: recorded observations and labels encoded as table cells or text. Their usefulness depends on consistent criteria, accurate transcription and enough context for the panel to interpret them properly.

    Before you try thisApplicant details and assessment notes are personal and HR-sensitive. They must not be pasted into an external or unapproved AI tool, because doing so could disclose personal information and compromise the selection process. Use only approved systems, restrict access and handle the file under the applicable retention and confidentiality instructions.

    Where you may hear it

    • A recruitment panel agreeing the fields in a selection matrix
    • A vendor demo to an ICT governance board
    • An inter-agency working group reviewing a needs assessment
    • A programme review before a monitoring report is cleared
    • A data protection briefing before access is granted to a new platform

    Avoid confusion

    Avoid confusion

    Often confused with

    Common misconceptions

    • Data only means numbers in a spreadsheet.

      Numbers are one kind of data, but text, images, audio, locations and timestamps can also be data if they are recorded and used for a purpose.

    • More data always gives better results.

      Quantity helps only if the data is relevant and of sufficient quality. Large amounts of outdated, biased or inaccurate data can make results worse.

    • Data and dataset mean the same thing.

      Data is a broad term for recorded information. A dataset usually means a defined collection of data assembled for a task such as analysis or model training.

    Quick check

    Which option best describes data in an AI context?

    Go deeper

    Go deeper

    The simple version

    Data is recorded information that a person or computer can use.

    The technical version

    Data consists of encoded observations, measurements or labels represented in a form that can be stored, processed and transmitted, such as rows in a table, text tokens, image pixels or sensor values.

    Reviewed external resources

    • Primary sourceRecommendation on the Ethics of Artificial IntelligenceUNESCO

      This link gives a policy and ethics perspective on how data should be handled in AI contexts, which suits readers concerned with rights, governance and responsible use.

      Intermediatelink checked 2026-07-31

    • EncyclopaediaDataWikipedia

      This link gives a broad overview of the term and its different uses, which suits readers who want a quick general starting point.

      Intermediatelink checked 2026-07-31

    • ArticleMachine Learning GlossaryGoogle

      This link gives short plain-language glossary definitions used in machine learning contexts, which suits beginners who want AI-specific vocabulary.

      Intermediatelink checked 2026-07-31

    • VideoVideo explanations: DataYouTube search

      This is a search results page for videos about data, which suits readers who prefer to compare several visual explanations before choosing one.

      search results, not one pagelink checked 2026-07-31

    • DiscussionCommunity discussion: DataReddit search

      This is a search results page for community discussions about data, which suits readers who want to see common questions, confusions and informal examples.

      search results, not one pagelink checked 2026-07-31

    Where this came from

    Drafted with a language model through an offline pipeline, checked against this site's schema, and every external link above fetched to see whether it answers. Treat it as a starting point and check anything that matters against the primary sources. How this site is built and what that means.

    The one thing to remember

    Data is recorded material in a usable form. What a system can show or learn depends on what was recorded, how it was encoded and how it was prepared.