GIS and Python for Agricultural Applications Training Course

Participants can organise agricultural spatial data using suitable formats, attributes and coordinate systems.

• Vector, raster and tabular structures used in agricultural information systems
• Field boundary schemas with crop, season, management and ownership attributes
• Coordinate reference systems selected for area, distance and local operations
• Geodatabases organised by source, season, processing stage and access controls
• Metadata records documenting source, date, resolution, limitations and licensing
• Joining survey tables to parcels using stable geographic identifiers

Participants can write and run basic Python scripts for agricultural data processing.

• Python environments, notebooks, scripts and reproducible project folder structures
• Variables, lists, dictionaries and functions for agricultural data tasks
• Pandas tables imported, inspected and cleaned from survey extracts
• GeoPandas GeoDataFrames created from field boundaries and observation points
• File paths and parameters separated from analytical processing logic
• Error messages interpreted through incremental testing and code inspection

Participants can clean, validate and derive useful attributes from farm and field datasets.

• Geometry validation identifies duplicates, gaps, overlaps and invalid field polygons
• Attribute cleaning standardises crop names, dates, units and missing values
• Field areas calculated in projected coordinate systems and checked
• Parcel identifiers reconciled across farm registers and seasonal survey files
• Spatial joins attach administrative, extension and agroecological classifications
• Python scripts generate cleaned layers and data-quality exception reports

Participants can design field observations and sampling plans that support agricultural spatial analysis.

• Observation protocols define crop stage, damage class and photo evidence
• Stratified random sampling across crop types, zones and management conditions
• Systematic sampling applied to irrigation schemes and demonstration plots
• Sample-size considerations for field validation and crop classification
• KoboToolbox forms structured for coordinates, repeat visits and validation rules
• Positional accuracy assessed against field size and intended decisions

Participants can select suitable satellite data for defined agricultural monitoring questions.

• Sentinel-2 spectral bands matched to crop and vegetation applications
• Landsat imagery compared by resolution, revisit interval and archive depth
• Cloud cover, shadows and atmospheric effects considered before analysis
• Scene selection aligned with planting, flowering, harvest and stress periods
• Spatial resolution matched to field size and management variability
• STAC catalogues used to identify accessible imagery and metadata

Participants can prepare satellite raster data for field-level agricultural analysis.

• Raster bands inspected for projection, extent, resolution and data type
• Area-of-interest clipping reduces processing volume and irrelevant pixels
• Cloud and shadow masks applied before spectral-index calculation
• Raster alignment ensures comparable pixels across dates and sensors
• NoData values handled before zonal statistics and temporal aggregation
• GDAL commands automate format conversion, clipping and raster compression

Participants can calculate and interpret spectral indices for crop and water-condition assessment.

• NDVI calculated from near-infrared and red reflectance bands
• NDWI distinguishes surface water and canopy moisture applications
• SAVI reduces soil-background influence in sparsely vegetated fields
• Index thresholds interpreted using crop stage, local conditions and validation
• Zonal statistics summarise pixel values within farm and field boundaries
• Python functions create index rasters and field-level indicator tables

Participants can analyse seasonal crop-condition patterns from repeated satellite observations.

• Time-series stacks assembled from cloud-screened seasonal image collections
• Field-level NDVI profiles compared across crops, locations and seasons
• Smoothing methods reduce noise from clouds and irregular observations
• Anomaly analysis compares current conditions against historical seasonal baselines
• Phenological metrics estimate green-up, peak growth and senescence timing
• Charts combine satellite indicators with rainfall and field-observation records

Participants can combine weather and climate variables with agricultural spatial datasets.

• CHIRPS rainfall grids extracted for fields, districts and catchments
• Growing degree days calculated from daily minimum and maximum temperatures
• Rainfall anomalies measured against seasonal climatological reference periods
• Dry-spell analysis identifies consecutive days below rainfall thresholds
• Weather-station observations compared with gridded climate-data estimates
• Spatial indicators linked to planting calendars and crop-stage assumptions

Participants can derive soil and terrain indicators relevant to agricultural land management.

• Digital elevation models produce slope, aspect and elevation derivatives
• Flow direction and accumulation identify drainage pathways and runoff concentration
• Terrain ruggedness flags mechanisation constraints and erosion-prone terrain
• Soil texture, pH and organic-carbon layers prepared for analysis
• Field observations compared against mapped soil-property predictions
• Reclassification converts continuous terrain and soil values into suitability classes

Participants can build transparent land suitability models for a specified crop or intervention.

• FAO land-evaluation principles connect land qualities with crop requirements
• Suitability criteria defined for rainfall, slope, soils and temperature
• Boolean constraints exclude protected areas, water bodies and unsuitable terrain
• Weighted overlay combines standardised criteria using documented assumptions
• Analytic Hierarchy Process supports pairwise weighting of suitability criteria
• Sensitivity analysis tests how weights and thresholds alter results

Participants can map irrigation demand and operational constraints using spatial agricultural data.

• FAO-56 reference evapotranspiration concepts applied to crop water requirements
• Crop coefficients matched to growth stages and cropping calendars
• Effective rainfall estimated for seasonal irrigation-demand calculations
• Irrigation command areas delineated from canals, parcels and elevation
• Water-demand summaries aggregated by scheme, block and delivery period
• Spatial constraints identify fields with drainage, access or elevation issues

Participants can develop and evaluate a supervised crop-classification workflow.

• Training samples prepared from field observations and verified parcel records
• Feature stacks combine spectral bands, indices and seasonal metrics
• Random Forest classification separates crop classes from reference data
• Training and validation samples kept separate to prevent optimistic accuracy
• Confusion matrices report producer, user and overall classification accuracy
• Class maps reviewed for mixed pixels, edge effects and implausible patterns

Participants can create cautious spatial estimates of yield and crop production from available data.

• Crop-cut data cleaned for plot area, moisture and unit consistency
• Yield observations linked to parcels, vegetation indices and weather variables
• Regression models relate yield measurements to explanatory spatial indicators
• Cross-validation assesses predictive performance and residual error patterns
• Production estimates combine mapped crop area with yield assumptions
• Uncertainty ranges communicated separately from headline production estimates

Participants can identify and interpret meaningful spatial patterns in agricultural datasets.

• Spatial autocorrelation tested using Global Moran's I statistics
• Local Moran's I identifies clusters and spatial outlier fields
• Kernel density estimation maps concentrated pest reports and service demand
• Spatial interpolation concepts distinguish measured values from estimated surfaces
• Cluster results checked against sampling density and reporting bias
• Maps avoid causal claims unsupported by observational agricultural data

Participants can automate recurring agricultural GIS workflows with reusable Python code.

• Functions package repeated clipping, indexing and zonal-statistics procedures
• Loops process multiple dates, districts, crops and seasonal datasets
• Configuration files store parameters, paths, thresholds and reporting periods
• Logging records processing steps, warnings and failed data inputs
• Assertions test expected fields, coordinate systems and value ranges
• Batch outputs named systematically for traceable reporting and archiving

Participants can communicate agricultural findings through decision-ready visual outputs.

• Map layouts apply scale bars, legends, sources and projection details
• Classification methods selected to avoid misleading agricultural choropleth maps
• Field maps distinguish observed data, model outputs and uncertainty
• Charts present seasonal indicators, comparisons and operational thresholds clearly
• Technical notes record methods, assumptions, limitations and update requirements
• Decision briefs connect geospatial findings with specific management actions

Participants can apply quality controls and governance practices to agricultural geospatial workflows.

• Quality-control checklists cover geometry, attributes, rasters and analytical outputs
• Version control principles protect scripts, parameters and documented changes
• Peer review tests assumptions, code logic and interpretation claims
• Reproducible workflows preserve inputs, environments, methods and output metadata
• Data-sharing protocols address farmer privacy, land rights and consent
• Operational handover plans assign update responsibilities and review schedules

Participants can build, review and present a complete agricultural geospatial analysis package addressing a defined decision problem.

• Decision question defined with users, scope, timing and success criteria
• Source datasets assessed for fitness, permissions, metadata and quality limitations
• Python and GIS workflow builds reproducible agricultural analytical outputs
• Satellite, weather, soil and field evidence integrated into findings
• Quality review checks calculations, maps, assumptions and uncertainty statements
• Technical presentation recommends actions, monitoring indicators and update procedures