This subsection covers the core data manipulation operations required to make spatial data from disparate sources compatible for combined use within a single GIS project — georeferencing (establishing real-world coordinates), format conversion (changing the file/encoding structure), and data transformation (changing the coordinate system or other fundamental data properties).

Georeferencing (introduced in subsection 401 in the context of manual digitizing) is the broader process of establishing the relationship between a dataset's internal coordinates and real-world geographic coordinates — required not only for scanned paper maps but for any imagery or dataset lacking pre-existing valid coordinate system information. Beyond the control-point-and-transformation approach already covered, georeferencing accuracy depends heavily on control point distribution (well-distributed points across the full extent of the image produce more reliable transformations than tightly clustered points) and the appropriate transformation order selection (a higher-order polynomial transformation can fit control points more closely but risks introducing unrealistic distortion in areas without nearby control points, especially near the image edges).

Format conversion changes the file structure/encoding used to store spatial data — converting a shapefile to a geodatabase feature class, converting GeoJSON to KML, or converting between different raster formats (GeoTIFF, IMG, JPEG2000). Format conversion is necessitated by the reality that different software platforms, web services, and organizational standards require different native formats, and almost no real-world GIS workflow operates exclusively within a single format throughout its full data lifecycle. Format conversion can be largely "lossless" (preserving all original geometry and attribute information, simply re-encoding it) for conversions between similarly-capable formats, but can also be "lossy" when converting to a format with more limited capabilities — for example, converting a geodatabase feature class with multiple attribute domains and subtypes to a basic shapefile will lose the domain/subtype enforcement rules, since shapefiles don't support that schema-level functionality.

Data transformation, in this specific context, most commonly refers to coordinate transformation/reprojection — converting a dataset's coordinates from one coordinate reference system to another, building directly on the datum and projection concepts from subsection 101. A coordinate transformation can involve a projection change alone (same datum, different projection — a purely mathematical re-calculation with no inherent accuracy loss beyond floating-point precision), a datum change alone (same projection, different reference ellipsoid/origin — requiring an appropriate datum transformation method, as discussed in subsection 101, since an incorrect or approximate datum transformation can introduce significant positional error), or both simultaneously.

Beyond coordinate transformation, "data transformation" in a broader data manipulation context can also encompass: schema transformation (restructuring attribute fields, renaming, recalculating, or reorganizing tabular data to match a target schema requirement), geometry type transformation (converting between related geometry types — for example, converting polygon features to their boundary line representation, or converting multipart features to single-part features), and generalization/simplification transformation (deliberately reducing the vertex density or detail level of features, often as part of preparing data for display at a smaller map scale, distinct from the involuntary generalization error discussed in manual digitizing).

A critical practical principle tested on the exam: every format conversion and coordinate/datum transformation should be evaluated for potential information loss or accuracy degradation BEFORE being applied to authoritative or legally significant data — maintaining an unmodified archival copy of the original source data alongside any transformed working copies is standard professional practice, since transformations are not always perfectly reversible (particularly lossy format conversions and certain datum transformation methods that rely on approximate regional grid corrections rather than exact mathematical formulas).

• Georeferencing: establishing real-world coordinates for data lacking valid existing spatial reference
• Lossless vs. lossy conversion: whether all original information is preserved or whether target format limitations cause information loss
• Projection-only transformation: mathematically exact, no inherent accuracy loss
• Datum transformation: can introduce meaningful positional error if using imprecise/approximate transformation methods
• Schema transformation: restructuring attribute fields/tables to match a target data model
• Archival best practice: retain unmodified original source data before applying transformations, since some are not perfectly reversible

Real-world application: A regional planning consortium integrating parcel data from five different counties — each maintaining data in a different local coordinate system, two different file formats (shapefile and geodatabase), and inconsistent attribute schemas — must systematically georeference any non-spatially-referenced legacy data, reproject all county datasets to a single common regional coordinate system, convert all data to a unified geodatabase format, and transform/standardize attribute schemas to a shared regional data model — while retaining each county's original unmodified source data in case any conversion step needs to be revisited or audited.

Common exam trap: Students sometimes assume all coordinate transformations are equally precise and risk-free. The exam expects you to distinguish between exact mathematical projection transformations (low risk) and approximate datum transformations using regional correction grids (higher risk of meaningful positional error if the wrong transformation method or an outdated/imprecise grid is selected for the specific region and datum pair involved).