GIS Data Acquisition & Sources: Methods, Formats & Real-World Data

TAKEAWAYS:

• GIS data is acquired via remote sensing (satellites, drones), field surveys (GPS, LiDAR), and crowdsourcing (mobile apps)—each with trade-offs in cost, accuracy, and timeliness.
• Vector data (points, lines, polygons) excels in precise boundaries (e.g., roads, property parcels), while raster data (grids) handles continuous surfaces (e.g., elevation, vegetation).
• TIN (Triangulated Irregular Network) balances detail and efficiency for terrain modeling, critical for hydrology and agriculture in Nepal.
• Open data portals (e.g., OpenStreetMap, Nepal GIS Portal) and commercial providers (e.g., Esri, Maxar) shape GIS applications—choose based on budget and legal constraints.
• Worked example: Daraz uses GIS to optimize delivery routes (vector network analysis) and predict demand (raster-based terrain analysis) in Kathmandu’s uneven terrain.
• Exam focus: Compare data structures (vector vs. raster vs. TIN), explain data acquisition methods, and link real-world Nepalese applications (e.g., Pathao’s dynamic traffic routing).

1. Introduction to GIS Data Acquisition

GIS relies on spatial data—information tied to Earth’s surface. Data acquisition is the process of collecting this data, which can be:

• Primary: Collected firsthand (e.g., GPS surveys, drone flights).
• Secondary: Obtained from existing sources (e.g., satellite images, census data).

Why Data Acquisition Matters

Without accurate data, GIS outputs (maps, analyses) are unreliable. For example:

• A flood prediction model in Nepal’s Terai relies on precise elevation data (from LiDAR or DEMs).
• Pathao’s ride-hailing app uses real-time traffic data (from GPS-enabled phones) to reroute drivers.

(Shows how satellites orbit Earth at ~700 km altitude, capturing multispectral imagery used for land cover classification.)

Key Data Acquisition Methods

Mermaid diagram of data acquisition workflow:

flowchart TD
    A["Data Need"] --> B["Primary Data"]
    B --> C["Field Surveys\n(GPS, LiDAR)"]
    B --> D["Remote Sensing\n(Satellites, Drones)"]
    A --> E["Secondary Data"]
    E --> F["Existing Databases\n(Govt., Commercial)"]
    E --> G["Crowdsourced Data\n(Mobile Apps)"]
    C --> H["Raw Data --> Processed Data"]
    D --> H
    F --> H
    G --> H
    H --> I["GIS Database"]

2. Primary Data Acquisition: Field Surveys and Remote Sensing

A. Remote Sensing: Eyes in the Sky

Remote sensing captures data without physical contact. Key platforms:

• Satellites: Orbit Earth at altitudes of 500–1,000 km, capturing visible, infrared, and radar data.
  • Example: Sentinel-2 (ESA) provides 10m resolution free imagery, used by Nepal’s Department of Forests to track illegal logging.
• Drones/UAVs: Fly at 100–500m altitude, ideal for small-scale projects.
  • Example: NTC’s road damage assessment after the 2015 earthquake used drone footage to map collapsed bridges.
• Aerial Photos: Taken from aircraft (e.g., Nepal’s Survey Department uses LiDAR-equipped planes to map Himalayan glaciers).

(Shows drone payloads like multispectral cameras and RTK modules for centimeter-level accuracy.)

Worked Example: Daraz’s Warehouse Layout Optimization Daraz uses LiDAR-scanned 3D models of its warehouses to:

1. Map aisle widths (vector data) for robot navigation.
2. Analyze storage density (raster heatmaps) to reduce dead space.
3. Simulate order-picking paths (network analysis) to cut delivery times by 20%.

Key Data Used:

• Vector: Wall outlines, shelf coordinates.
• Raster: Floor reflectance (for robot vision).
• TIN: Ceiling height variations (to avoid collisions).

B. Field Surveys: Ground Truth

For high-precision data, GIS teams use:

• GPS (Global Positioning System): Provides horizontal accuracy ±2m (standard GPS) or ±2cm (RTK-GPS).
  • Example: Ncell’s tower sites are surveyed with RTK-GPS to ensure signal coverage in Kathmandu’s valleys.
• LiDAR (Light Detection and Ranging): Lasers measure 3D terrain in millions of points per second.
  • Example: Nepal’s hydropower projects (e.g., West Seti) use LiDAR to model riverbeds for dam design.
• Total Stations: Combine electronic theodolites + EDM for surveying.
  • Example: Property boundary disputes in Kathmandu’s urban areas are resolved using total station data.

(Shows point clouds in green/yellow/red, representing tree heights and gaps.)

3. Secondary Data Acquisition: Leveraging Existing Sources

Not all data needs to be collected fresh. Secondary sources include:

• Government Portals:
  • Nepal GIS Portal (topography, land use).
  • Central Bureau of Statistics (CBS) (population, census data).
• Commercial Vendors:
  • Esri’s ArcGIS Online: Subscription-based, includes global basemaps and historical imagery.
  • Maxar Technologies: High-resolution satellite imagery (used by Nepal Police for disaster response).
• Open Data Initiatives:
  • OpenStreetMap (OSM): Crowdsourced maps (e.g., Pathao’s traffic data is contributed by drivers).
  • NASA’s EarthData: Free satellite data (e.g., MODIS for Nepal’s air quality monitoring).

Comparison Table: Secondary Data Sources

4. GIS Data Structures: How Data is Stored

Data structures define how spatial data is organized in GIS. The three primary types are:

A. Vector Data: Points, Lines, and Polygons

• Definition: Stores data as geometric objects (points, lines, polygons) with x,y coordinates.
• Example:
  • Points: Tree locations in a forest.
  • Lines: Roads, rivers.
  • Polygons: Administrative boundaries (e.g., Kathmandu Metropolitan City limits).

(Shows how each feature has attributes like length, area, or ID.)

Advantages:

• Precise for discrete features (e.g., property lines).
• Supports topological relationships (e.g., "Road A connects to Road B").

Disadvantages:

• Poor for continuous surfaces (e.g., elevation, temperature).
• Requires more storage for complex polygons.

Nepalese Application:

• Ncell’s network planning: Vector data models cell tower locations and signal coverage polygons.

B. Raster Data: Grids of Cells

• Definition: Divides Earth’s surface into a grid of squares (pixels), each with a value (e.g., elevation, vegetation index).
• Example:
  • Digital Elevation Model (DEM): Each cell’s value = elevation (e.g., 100m, 200m).
  • Land Cover Map: Each cell’s value = land type (forest, urban, water).

(Shows color-coded elevation layers, with a legend indicating meters above sea level.)

Advantages:

• Ideal for continuous surfaces (e.g., terrain, climate).
• Supports fractional values (e.g., 70% forest cover).

Disadvantages:

• Low precision for small features (e.g., a single tree).
• Data redundancy (adjacent cells may have similar values).

Nepalese Application:

• Nepal’s hydropower dams: DEMs predict flood zones and water flow paths.

C. TIN (Triangulated Irregular Network): Hybrid Approach

• Definition: Connects irregularly spaced points with triangles to model terrain efficiently.
• Example: Glacier surface mapping in the Himalayas (where elevation changes rapidly).

Mermaid diagram of TIN vs. Raster vs. Vector:

classDiagram
    class Vector {
        +Points, Lines, Polygons
        +Topological relationships
    }
    class Raster {
        +Grid cells
        +Continuous surfaces
    }
    class TIN {
        +Triangles from irregular points
        +Balances detail & storage
    }
    Vector --> "Uses" TIN : "For complex boundaries"
    Raster --> "Complements" TIN : "For smooth surfaces"

Advantages:

• Less storage than raster for complex terrain.
• More accurate than raster for sharp features (e.g., cliffs).

Disadvantages:

• Complex to process (requires triangulation algorithms).
• Not ideal for small-scale analysis.

Nepalese Application:

• Nepal’s disaster risk mapping: TIN models landslide-prone slopes in the Mid-Hills.

5. Data Quality and Validation

GIS data is only useful if accurate and reliable. Key quality checks:

1. Positional Accuracy: How close data is to "ground truth" (e.g., GPS error ±2m).
   • Test: Compare field-surveyed points with satellite data.
2. Attribute Accuracy: Correctness of labels (e.g., "forest" vs. "urban").
   • Test: Cross-check with aerial photos.
3. Completeness: Are all features included? (e.g., missing roads in OSM).
   • Test: Use crowdsourced edits (e.g., Pathao drivers reporting new roads).
4. Consistency: No contradictions (e.g., a river flowing uphill).

(Shows how RTK-GPS points cluster within 2cm, while standard GPS scatters by ±2m.)

6. Data Sharing and Standards

GIS data must follow standards to ensure compatibility. Key formats:

Data Sharing Platforms:

• Open Data Portals: data.gov.np (Nepal’s open data hub).
• ArcGIS Online: Esri’s cloud-based sharing.
• Figshare: Academic data repository.

In the Real World

1. Pathao’s Dynamic Routing

   • Idea Used: Vector network analysis (shortest path algorithms) + raster terrain data (elevation to avoid steep hills).
   • How: Pathao’s app uses OpenStreetMap vector data for roads and DEMs to reroute drivers away from landslide-prone areas in Kathmandu’s valleys. During the 2015 earthquake, crowdsourced data (e.g., blocked roads) updated the system in real time.

2. Ncell’s 5G Tower Placement

   • Idea Used: LiDAR TIN models for terrain + vector building footprints.
   • How: Ncell’s engineers use Esri’s ArcGIS to overlay high-resolution LiDAR data (from drone surveys) with urban building polygons (from satellite imagery). This helps place towers in high-traffic areas while avoiding signal-blocking structures like mountains or dense forests.

3. Daraz’s Inventory Optimization

   • Idea Used: Raster heatmaps (storage density) + vector aisle layouts.
   • How: Daraz’s warehouses in Kathmandu and Pokhara use 3D LiDAR scans to create TIN models of warehouse interiors. Raster analysis identifies underused storage zones, while vector data maps robot navigation paths, reducing order fulfillment time by 30%.

Exam Tip

This unit tests three core areas:

1. Data Acquisition Methods (5–10 marks):

   • Compare remote sensing vs. field surveys (e.g., "Why does Nepal’s hydropower sector prefer LiDAR over drones for dam sites?").
   • Discuss crowdsourcing challenges (e.g., "How does OpenStreetMap’s data quality affect Pathao’s navigation?").
   • Worked example: Describe how Ncell uses GPS + LiDAR to plan 5G towers in the Himalayas.

2. Data Structures (10–15 marks):

   • Compare vector, raster, and TIN in a table (include advantages, disadvantages, and Nepalese examples).
     • Example answer snippet:

       Feature	Vector	Raster	TIN
       Best for	Roads, property lines	Elevation, land cover	Complex terrain
       Storage	Low (for simple shapes)	High (grid redundancy)	Medium (triangles)
       Nepal Use	Ncell’s cell towers	DEMs for hydropower	Glacier mapping

   • Explain why raster is used for flood modeling but vector for road networks.

3. Data Sources and Standards (5 marks):

   • Short-answer: "Why does Nepal’s government prefer OpenStreetMap for disaster response over commercial data?"
     • Answer: Cost (free vs. $), real-time updates (crowdsourced), and local relevance (drivers report road blocks).
   • Long-answer: Describe how Daraz combines OSM vector data with DEM raster data to optimize deliveries in Kathmandu’s uneven terrain.

Common Pitfalls:

• Ignoring Nepal’s context: Always tie examples to Himalayan terrain, urban sprawl, or disaster risks.
• Overlooking data quality: Exams ask about validation methods—mention GPS accuracy tests or cross-checking with aerial photos.
• Vague comparisons: When contrasting vector vs. raster, use specific Nepalese cases (e.g., "Ncell uses vector for towers but raster for signal coverage maps").

Final Tip: For 10-mark questions, structure your answer with:

1. Definition (1 mark).
2. Comparison table (4–6 marks).
3. Nepalese example (2–3 marks).
4. Real-world impact (1 mark).

For 5-mark questions, focus on:

• Methodology (e.g., "How does LiDAR work?").
• Application (e.g., "Why is DEM critical for Nepal’s hydropower?").
• Challenge (e.g., "What’s the biggest limitation of crowdsourced GIS data?").

Mermaid diagram of exam question breakdown:

flowchart TD
    A["10-Mark Question"] --> B["Define\n(1 mark)"]
    A --> C["Compare\n(6 marks)\n(Table + Nepal example)"]
    A --> D["Impact\n(3 marks)"]
    E["5-Mark Question"] --> F["Method\n(2 marks)"]
    E --> G["Application\n(2 marks)"]
    E --> H["Challenge\n(1 mark)"]