Introduction to GIS Methods in Economics

Giorgio Chiovelli Sebastian Hohmann

Bonn, 28/05/2019
Overview
The plan for today

Digitization of Old Maps

Geo-reference
Digitization
NetCDF data
Using NetCDF in ArcGIS
Interpolation Methods
Interpolation
Example: Create Time Series of City Temperature using Interpolation
Further Data Visualization Tools
Hot Spot Analysis
Heat Map
Geocoding
Geocoding through API
Maps
Digitizing an old map

From scanned maps to digitized maps

Digitize a paper map can be rewarding
- Create new units of observation (e.g. Murdock’s Map)
- Time variation (e.g. Time series of Road construction in Burgess et al.
2015)

Unfortunately, this is one of those tasks that cannot be automized in ArcGIS

In this section we will show you how to create shapefiles from a paper map
Maps
Import a scanned map in ArcMap

Tips before importing the map

Always choose a high resolution format
- Zoom as much as possible your pdf, copy and paste in Paint/Photoshop,
save as .tif

Knowing the coordinate system of the original map would help reducing
distortion during georeferencing

Import your map clicking the , browse to your picture and import without double
clicking on it.
Maps
Georeferencing Toolbar

How to assign projection to your map

After importing it, your map is now floating around your working environment

We need to pin down the map to using the georeferencing toolbar

(if you do not see the toolbar, right click on the command toolbar and check
georeferencing)
Maps
Georeferencing Toolbar

Interface
Maps
How to assign projection to your map

Functions you are going to use

Update Georeferencing: Saves the transformation with the raster.

Rectify: Creates a new transformed raster dataset.

Fit To Display: Shifts the raster to the current display extent area.

Transformation
Maps
How to assign projection to your map

Functions you are going to use

Add Control Points : Allows you to select control points from a layer and add
them to the map.
- Do not start by georeferencing too many points that are close to each
other (spatially correlated)
- First three points are the most important
- Once you have several links you can try to experiment different
Transformations (the more links you have the best it is)
- If possible, use geographical features that already exist (roads, cities,
rivers, etc.)

View Link Rectify: Shows links and errors in tabular form.

Maps
Georeferencing Toolbar

Final output
Maps
Digitizing you georeferenced map

Create a digital version of you map

Now we have a georeferenced map that we can transform into a shapefile

Task: we want to create a shapefile of roads at colonial time (1973) in

Mozambique

What should we use?

Maps
Digitizing you georeferenced map

Prepare your ground

Add all the necessary layers you want to work with in the Map Project to
ArcMap.
- e.g. 2011 shapefile of roads in Mozambique

Add the new, empty Shapefile Layer you created in ArcCatalog to which you
want to add Features.
- In our case we would create an empty line shapefile
Add the Editor Toolbar to Your Tools
Maps
Digitizing you georeferenced map

Prepare your ground

Create Feature : open the Create Features window so that you can create
new features
Maps
Digitizing you georeferenced map

Start Drawing
Straight Segment : create a vertex each time you click.

Trace : create segment by following existing features.

- Very useful in case you have underlying feature that overlaps with your
map
Finish a sketch: right click on last vertex, click on Finish Sketch . Or short
cut: press F2.
Attribute : opens attribute window so you can modify the attribute values of
the selected feature
- Very important to fill in information on the feature (e.g. road condition,
paved/unpaved)
Maps
Digitizing you georeferenced map

Final shapefile
NetCDF Data

Network Common Data Form

File format for storing multidimensional scientific data
e.g. Temperature, Humidity, Pressure, Wind speed etc.
Each variable can be displayed through a dimension [like Time]

NetCDF in ArcGIS
Unpack the NetCDF
Create raster from NetCDF
Generate tables from NetCDF
Working with NetCDF Data
Multidimension Tools

What tool to use?

Multidimension Tools
Working with NetCDF Data
Data

Example of NetCDF data

https://www.metoffice.gov.uk/hadobs/hadcrut4/data/4.5.0.0/
download.html
Download HadCRUT.4.5.0.0.median netcdf.zip
Working with NetCDF Data
Importing Data

You cannot add a netCDF file using the button

How to import NetCDF data?

Working with NetCDF Data
Displaying different time values

load a world map shapefile to see what is where

right click on temperature anomaly Layer1 → Properties...
select the NetCDF tab
under Dimension Values, click on Value and select the point in
time from the drop down menu you are interested in
Interpolation
Interpolation Analysis

Why interpolating?
To predicts values for cells in a raster from a limited number of sample
data points.

It can be used to predict unknown values for any geographic point data

elevation, rainfall, pollution, etc

Assumption
Good option if objects/points are spatially correlated
Interpolation
Example: Elevation

What is interpolation?
Interpolation
Example: Rainfall

What is interpolation?
Interpolation
Example: Ozone Concentration

What is interpolation?
Interpolation
Interpolation Methods

To access the following tools, you need to authorize the Geostatistical Analyst
license.

Different Methods of Interpolation

Many methods exist

Check the following link for a complete list of options

http://desktop.arcgis.com/en/arcmap/10.6/extensions/
geostatistical-analyst/
an-introduction-to-interpolation-methods.htm

Appropriate method is application-specific

Interpolation
Inverse Distance Weighted (IDW) Interpolation

Idea
“Objects close to one another are more likely to be alike”

IDW uses the measured values surrounding the prediction location

IDW assumes that each measured point has a local influence that
diminishes with distance

greater weights to points closest to the prediction location

Interpolation
Inverse Distance Weighted (IDW) Interpolation
Paper examples: temperature interpolation
Temperature Interpolation. Waldinger, 2017

Maria Waldinger (2017) “The Economic Effects of Long-Term Climate Change

- Evidence from the Little Ice Age,”.

Motivation
Literature on effects of short-term climate variation on economic outcomes but
little evidence on long-term effects of climate change.

Contribution
Using Little Ice Age in 1500 as natural experiment. Combining annual
temperature data and population counts to generate a panel data set for 2120
European cities. Negative effect of climate shock on city size, decreased crop
yields, and increased crop prices. Effect mitigated by access to trade.
Paper examples: temperature interpolation
Temperature Interpolation. Waldinger, 2017
Temperature Reconstruction for Moscow 1520-1750

How can we replicate a similar dataset for temperature at city-level?

Example of Loop in Python
Create Time Series of City Temperature using Interpolation

Objective
Want to create annual mean temperature deviations for European capital
cities

Show how to
- import data from text file,
- filling up missing value using IDW interpolation
- extract raster value to points

Iterate from 1985 to 2007

Script: interpolation example.py

Example of Loop in Python
Create Time Series of City Temperature using Interpolation

Data
European 1400 Year Spring-Summer Temperature Reconstructions (Guiot
et al. 2012 data)

Data reconstructing Temperature deviations for 600-2007.

117 proxy records, including tree-rings, documentaries, pollen
assemblages, and ice cores. Grid size 5°x5°. Data are °C anomalies
relative to the 1961–1990 average.

https://www.ncdc.noaa.gov/paleo-search/study/10426
Example of Loop in Python
Create Time Series of City Temperature using Interpolation

Observation Points
Example of Loop in Python
Create Time Series of City Temperature using Interpolation

IDW Raster
Example of Loop in Python
Create Time Series of City Temperature using Interpolation

IDW Raster - Search Neighborhood: Standard

Example of Loop in Python
Create Time Series of City Temperature using Interpolation

IDW Raster - Search Neighborhood: Smooth

Data Visualization Tools

How to visualize Spatial Pattern in your data

Heat Map
Hot Spot Maps

Advantages
Help you detecting spatial patterns: cluster, spatial correlations
Appreciated by policy-makers and practitioners
Data Visualization Tools
Heat Map

Heat Map
Represents the geographic density of features on a map

Useful for layers with a large number of features

Heat Map in ArcGIS

Point Density: calculates the density of point features around each raster
cell.

Line Density: calculates the density of line features within a radius

around each output cell.

Kernel Density: algorithm to calculate density of point features around

each neighborhood. The algorithm determines the default search radius
(bandwidth), which allows for better weighting of highly dense points and
smoother outputs.
Data Visualization Tools
Heat Map, data

Data source
https://data.police.uk/data/
dowload the data for Surrey, March 2019
force boundaries are from here:
https://data.police.uk/data/boundaries/
We have treated the data for you and assigned a weight of 10 to all
violent and sexual crime, 5 to Criminal damage and arson, and 1 to
all other type of crime (theft, drugs etc.)
Data Visualization Tools
Heat Map

Point Density
Data Visualization Tools
Heat Map

Kernel Density
Data Visualization Tools
Hot Spot Analysis

Hot Spot Analysis

Identifies statistically significant spatial clusters of high values (hot spots)
and low values (cold spots).

Indicate whether the observed spatial clustering of high or low values is

more pronounced than one would expect in a random distribution of
those same values.
Data Visualization Tools
Hot Spot Analysis

Optimized Hot Spot Analysis

Geocoding outside GIS

Geocoding through API

Imagine you have a list of city and villages to which you want to attach
the corresponding GPS coordinates
We present a STATA script scraping information from
http://www.geonames.org
- See here to register for a free account (need to enable free web
services)
http://www.geonames.org/export/web-services.html
- Additional dofile geocoding via Google Maps (give you free
budget of $300/year, after using this, need to buy credit,
requires credit card to sign up, but will not charge unless you
explicitly ask it to)

Requires installation of libjson, insheetjson packages in STATA

Let us have a look at the dofile