ECON 0150 | Economic Data Analysis

The economist’s data analysis skillset.

Part 1.6 | Relationships In Space

Geographic Data

Some types of relationships in space

Geographic data is data organized on three axes (latitude, longitude, altitude)
We typically only use latitude and longitude
Geo data is often combined with other variables like population
Two main types of geo data: points, shapes
We sometimes observe points, but most data comes in groups

Example: Restaurants by Zipcode

Are there fewer restaurants further from downtown Pittsburgh?

We’re going to use a Census maps and openly available data on restaurant locations to answer this question.

Data: Census Shapefiles and Open Street Maps

Geographic Data

Maps are (typically) plots on two axis

> a basic map of Pittsburgh

Geographic Data

Maps can show any level of detail available in the data

> a map of Pittsburgh Zipcodes

Geographic Data

We can add information: colors

> a map of Pittsburgh Zipcode populations

Geographic Data

We can add information: colors

> a map of Pittsburgh Zipcode populations: interactive!

Geographic Data

Maps can also show points

> some restaurants in Pittsburgh!

Geographic Data

Maps can also show points

Geographic Data

Maps can also show points

Points are nice but we typically can’t use them raw
Some point transformations: distances between points; group by area; etc
We can also relate points to other variables (eg. zipcode population for each restaurant)

Geographic Data Example: Nunn (2008)

Did the historical trade of enslaved people impact modern economic development in Africa?

Method: Uses historical data and the distance from major ports

Findings: Areas more disrupted by enslavement have lower GDP today, due to:

Weakened institutions (political fragmentation, mistrust).
Disrupted societies (population loss, economic stagnation).

Implication: Historical shocks can have persistent economic effects.

Geographic Data Example: Weidman (2024)

Does the party of your neighbors impact your decision to vote?

Geographic Data Example: Weidman (2024)

My dissertation involved measuring distances between voters

Geographic Data

Are there fewer restaurants further from downtown Pittsburgh?

> lets get back to our question!

Geographic Data

Are there fewer restaurants further from downtown Pittsburgh?

Geographic Data

Are there fewer restaurants further from downtown Pittsburgh?

Steps:

Measure points by zipcode area

Measure distances between groups (take the centroid, etc)

Geographic Data

Subquestion 1: how many restaurants are in each Pittsburgh zipcode?

Geographic Data

Subquestion 2: how far is each zipcode from downtown?

> measure from the center (centroid) of the zipcode

Geographic Data

Subquestion 2: how far is each zipcode from downtown?

> measure from the center (centroid) of the zipcode

Geographic Data

Subquestion 2: how far is each zipcode from downtown?

> what’s the distribution?

Geographic Data

Subquestion 2: how far is each zipcode from downtown?

> we now have enough to answer our main question!

Geographic Data

Are there fewer restaurants in areas further from downtown Pittsburgh?

Homework 1.6 | US City Population and Temp

We’re going to use data on locations (lat, lng), population, and temperature (avg_temp) of US cities to map temerature and examine whether there is a relationship between latitute (north/south) is related to temperature.

Data: US_Cities.csv and Eastern_Cities.csv

Part 1 Wrap Up