\[ \newcommand\si[1]{\mathrm{#1}} \newcommand\SI[2]{#1\,\si{#2}} \newcommand\matr[1]{\mathbf{#1}} \DeclareMathOperator*{\argmax}{arg\,max} \DeclareMathOperator*{\argmin}{arg\,min} \]

# GIS data models

## 1 GIS data

Data: observations made from monitoring the real world

Information: data with meaning and context added

Wide variety of data sources

- National Elevation Dataset
- The National Map
- National Land Cover Database
- Soil Data
- And many more...

## 2 Data dimensions

All GIS data have three modes or dimensions:

- Spatial: values or symbols that convey information about the location of observed features
- Temporal: when data was collected
- Thematic: describes the characters of real-world features to which data refers, referred to as attributes

Need to be able to identify all three modes of data.

Data may be organized by any dimension.

GIS data must have a mathematical spatial reference called a coordinate system to locate the position of the feature.

### 2.1 Gainesville, GA tornado

- Spatial: Gainesville, GA, Downtown Square. Tornado track: 2.5 miles long, 0.5 mile wide
- Temporal: April 6, 1936 08:45
- Thematic: damage created by F4 tornado triggered by two thunderstorm cells that merged over downtown Gainesville in Hall County, GA

## 3 Data models

GIS is a model of reality.

GIS models include

- Spatial forms
- Spatial processes

## 4 Spatial entities

- Points, lines, and polygons
- Networks and surfaces

Network is a series of connecting lines along which things flow.

Surface entities represent **continuous** features (e.g., elevation, temperature).

Networks, points, lines, and polygons represent **discrete** data.

## 5 Modeling surfaces

Digital Terrain Models (DTM) approximate a continuous surface using a finite number of observations.

- Raster DTM
- Vector DTM

### 5.1 Raster DTM

Grid of height values (one per cell)

Accuracy depends on the complexity of terrain and resolution.

Digital Elevation Model (DEM)

### 5.2 Vector DTM

Triangulated Irregular Network (TIN)

Esri Terrain Dataset

LiDAR (Light Detection and Ranging) based

## 6 Networks

Lines: network links (roads, pipes, rivers)

Points

- Nodes: end of network link (junctions, valves, confluences)
- Stops: locations that may be visited or where transfer occurs to the system (e.g., bus stop, sediment sources)
- Centers: resource supply or attraction (e.g., airports, hospitals, malls)

Turns: transition from one link to another

### 6.1 Network properties

Impedance: cost associated with traversing a network link, stopping, turning, or visiting a center (e.g., travel time, fuel, driverâ€™s pay)

Supply and demand

- Supply: quantity of a resource available at a center (e.g., number of hospital beds available)
- Demand: utilization of a resource by an entity associated with a network link or node (e.g., number of people requiring treatment)

### 6.2 Network topology

Correct topology and connectivity is important.

Correct geography is not vital.

Impedance and distance should be preserved.

## 7 Homework: Global spatial autocorrelation test of the median income

**Keywords:**global Moran’s $I$, $p$-value

Use the median income layer from the book’s exercise 3b (IL_med_income.shp) to calculate the global Moran’s $I$ (a global measure of spatial autocorrelation) of the median income. Does the median income have any spatial pattern or not? Report the index and $p$-value. Write a short report with a screenshot, Moran’s $I$, and the p-value. Discuss your findings from this analysis. Draw your conclusion statistically using the $p$-value. Please upload the report in the PDF format to D2L.

**Hints**

- Run Geoprocessing → Spatial Autocorrelation (Global Moran’s I) on the median income layer
- Find the index and $p$-value from the output window; this tool won’t create any layer outputs
- If $p\le 0.05$, we’re 95% confident that the median income has spatial autocorrelation (a spatial pattern)
- Otherwise, the median income has no spatial patterns