01
Data
Satellite and field data
Land Sensing Geospatial Analytics
From Satellite to Decision
Satellite observation, machine learning and field verification, combined into evidence that you can act on.
The four stages at a glance
Four stages, every project
Every project follows the same transparent process: select suitable data, analyse it, check results against field evidence, and deliver clear outputs that support decisions.
02
Analysis
Mapping and change detection
03
Field verification
Reference points and accuracy checks
04
Decision
Maps, statistics and reports
1
Stage 1
Data — what we observe
We select data based on the question, area size, required detail, available budget and cloud conditions.
Satellite imagery (free)
Useful for regional coverage and repeated monitoring.
- Sentinel-2 (optical)
- Landsat (optical)
- MODIS (optical)
- Sentinel-1 (radar)
Covers large areas and supports comparisons over time.
High-resolution imagery
Commercial or licensed imagery when small features need more detail.
Used selectively where extra detail changes the analysis.
Other geospatial data
- Elevation models (DEM)
- Rainfall and temperature records
- Existing maps and boundaries
- Administrative and watershed data
Drone and field data
Local observations that add detail and help check satellite interpretation.
- GPS reference points
- Geotagged photos
- Short field observations
Free data first: use open satellite data where it is suitable. Licensed imagery or field collection should be proposed only when the project needs it.
2
Stage 2
Analysis — turning data into information
We combine remote sensing, geospatial processing and statistical methods to produce interpretable results.
Land and crop mapping
Classify land use, crop types and seasonal patterns using satellite imagery and reference data.
- Land use and land cover
- Crop types and seasonal change
- Area estimation and class statistics
Spectral and climate indices
Use indicators such as NDVI for vegetation, NDSI for snow and SPI for drought when relevant to the question.
NDVI · VegetationNDSI · SnowSPI · DroughtLST · Temperature
Time series and change detection
Compare seasons or years to detect change in crops, vegetation, water or land cover.
HighIndexLow
Illustrative spring cropIllustrative summer crop
Cloud processing and reproducible methods
Where appropriate, cloud platforms can process large satellite collections and keep a clear record of the method.
Satellite data→Processing→Validated map
3
Stage 3
Field verification — checking results on the ground
Field evidence helps test whether the map reflects real conditions. The sampling design and limits are reported with the results.
Field data collection
- Collect GPS points at selected locations
- Record photos and brief observations
- Cover relevant land and crop classes
Reference pointsVerified sample locations
↓
Training dataUsed to build the model
Validation dataHeld back for testing
↓
Accuracy assessmentConfusion matrix, class metrics and limitations
The proportion of training and validation samples depends on the project design. Any percentages or accuracy figures on the website must match the real project report.
4
Stage 4
Decision — clear results you can use
Outputs are prepared for the people who need to plan, report, allocate resources or monitor change.
Maps
Clear, georeferenced maps for viewing and print.
GeoTIFF · PDFGIS data
Spatial layers ready to use in GIS software.
GeoTIFF · ShapefileArea statistics
Area estimates, summary tables and class totals.
Excel · CSVAccuracy report
Methods, validation and result limitations.
PDFDashboards
Interactive maps and key project indicators where needed.
Web dashboardPlain-language reports
Concise findings and implications for decisions.
PDFWorked example
Khulm crop mapping — from data to decision
An example of how a crop mapping project may progress through the same four stages. Replace this example with approved project details and verified results before publication.
Choose suitable satellite observations and define the study area.
→
Prepare seasonal composites and classify the relevant crop types.
→
Compare mapped classes with reference points and field observations.
→
Deliver validated maps, area estimates and a clear methods summary.
Tools we use
Established geospatial tools
Google Earth EngineCloud processing
GIS softwareSpatial analysis and mapping
RStatistical analysis
PythonData processing and modelling