Viren Bajaj,
Dylan Steele,
Can Bostanci.
Some regions do not have systems in place to conduct economic surveys or other means of collecting data on their financial situations. In our project we attempt to address this problem by using publicly available satellite images to predict the wealth of a city (or, more generally, a geographic region) based on fundamental features identified in these images and running them through a support vector machine (SVM).
- Make sure you have the following in same directory as the notebook: a. night_images #(with the images in this directory) b. config_secret.json #(Google API key stored as JSON)
- Have OpenCV installed. Use
brew install homebrew/science/opencv
(on a Mac) to do so. Otherwise, visit http://opencv.org.
NBViewer Link: http://nbviewer.jupyter.org/github/electsigon/The-Wealth-of-Cities/blob/master/The%20Wealth%20of%20Cities.ipynb Link to night_images: https://drive.google.com/drive/folders/0B2kckEKTUMfjQnI5YWVIM0RoUEU?usp=sharing Anyone with an Andrew email can view this folder and download it.
Sources: http://city-data.com Google Static Maps API