Introduction to AI and Decision Making
B.Sc course, University of Debrecen, Department of Data Science and Visualization, 2024
Data types
Structured data
- Tabular data: Data stored in databases, tables, with well-defined rows and columns, such as relational databases (SQL).
- Time series data: Measurement results stored in a structured form, where the values assigned to different times are present (e.g. stock prices, changes in sensor data over time).
- Network data (Graph data): Networks of connections, such as social media networks or route networks. - - Although in some cases it can be considered semi-structured, they often have a well-defined structure.
Semi-structured data
- Document-based data: data in XML, JSON format, which are structured at a certain level, but not in a rigid tabular format.
- Spatial (geospatial) data: GPS coordinates, map data. These are usually stored in a semi-structured format (e.g. GeoJSON, KML).
- Sensor data: Although structured to some extent, it is often present in semi-structured form, especially when different types of sensors send different data or when different formats (e.g. CSV or JSON) are used.
Unstructured data
- Text data: Content of articles, e-mails, chat messages, web pages. These are often completely unstructured and do not have a well-defined format.
- Audio data: Voice recordings, music, sound effects, which usually do not contain a structure (e.g. MP3, WAV).
- Video data: Movies, moving image content (e.g. MP4, AVI), where the structure is minimal.
- Image data: Photos, images (e.g. JPG, PNG), which are usually stored in an unstructured form.
- Holographic and 3D data: 3D models, CAD files that usually do not have a standard structure (eg STL, OBJ).
- Binary data: Data that is stored in raw binary form and has no explicit structure (e.g. image files, program code).