Spatial: Networks, Interaction, & Econometrics for Migration
NERSA/NARSC Program in Mobility, Migration & Regional Science
1 Install and Run
To run locally:
- install pixi
- clone this repository, follow below
To run in codespaces:
- click the ‘open in codespaces’ button above
- the default config will include at least 8 cores (dont go below 32gb memory), but you can crank it up for better performance
pixi installpixi run jupyter lab
if running in a codespace this starts a jupyter server on the VM’s ‘localhost’. Open an existing notebook (or start a new one). When selecting a kernel, choose ‘Existing jupyter server’ and copy/paste the URL with token from the terminal
edit the notebooks as you like
if you want to use the jupyterlab interface instead of VSCode:
- in the bottom panel click the “ports” tab
- click the “add port” button
- type 8888
ports
- click back in the vscode interface
- click the “terminal” tab in the bottom panel
- scroll up and copy the token string from the URL
token - click back into jupyterlab and paste the token
2 Schedule
2.1 Overview
- spatial data and analysis in python
- encoding spatial relationships
- model flows via spatial interaction
- model flows via spatial econometric interaction
2.2 Introduction
open source and open science
- pixi & conda
- jupyter
- python
- python
- pysal
- geosnap
2.3 spatial data in python
- pandas
- numpy
- shapely
- geopandas
- census and lodes
2.4 spatial graphs
encoding relationships between spatial objects
- contiguity
- k-nearest neighbors
- distance
- flows
2.5 spatial interaction models
modeling the flow between \(o\) and \(d\) as a function of push and pull factors
- gravity model (unconstrained)
- attraction constrained
- production constrained
- doubly constrained
2.6 spatial econometric models
multiple graphs: modeling spatial dependence in the flow between \(o\) and \(d\) via \(^oW\), \(^dW\), and \(^{od}W\)
- spatial dependence
- OD-Graphs
- lag models
- error models