Tuesday, May 22, 2018

Using deep learning and satellite imagery to improve land use classification in cities

Marta Gonzalez and colleagues have a recent paper using deep learning and satellite image data to improve land use classification. The authors have made documented code and Jupyter notebooks available hereI'm self recommitting the paper and code to my future self. HT Marco De Nadai.

Albert, A., Kaur, J., & Gonzalez, M. C. (2017, August). Using convolutional networks and satellite imagery to identify patterns in urban environments at a large scale. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1357-1366). ACM.

Urban planning applications (energy audits, investment, etc.) require an understanding of built infrastructure and its environment, i.e., both low-level, physical features (amount of vegetation, building area and geometry etc.), as well as higher-level concepts such as land use classes (which encode expert understanding of socio-economic end uses). This kind of data is expensive and labor-intensive to obtain, which limits its availability (particularly in developing countries). We analyze patterns in land use in urban neighborhoods using large-scale satellite imagery data (which is available worldwide from third-party providers) and state-of-the-art computer vision techniques based on deep convolutional neural networks. For supervision, given the limited availability of standard benchmarks for remote-sensing data, we obtain ground truth land use class labels carefully sampled from open-source surveys, in particular the Urban Atlas land classification dataset of $20$ land use classes across $~300$ European cities. We use this data to train and compare deep architectures which have recently shown good performance on standard computer vision tasks (image classification and segmentation), including on geospatial data. Furthermore, we show that the deep representations extracted from satellite imagery of urban environments can be used to compare neighborhoods across several cities. We make our dataset available for other machine learning researchers to use for remote-sensing applications.

Tuesday, May 15, 2018

Urban Picture

Nova Iguaçu (Brazil). Nova Iguaçu is a municipality in the Metropolitan Area of Rio de Janeiro. The ocean and some of Rio's mountains can be seen in the background of the picture.

source: vonsenke on reddit, HT Vitor Gabriel

Sunday, May 13, 2018

Against All Authority

Happy mothers day

credit: ?

Tuesday, May 8, 2018

Visualizing space-time networks

I've said this on Twitter before but I should say it here as well. Craig Taylor and the Ito World team have some of the best data visualizations of geospatial data related to cities and transport networks.

Just a few days ago, Craig tweeted some of his latest work with neat visualizations of drive-time network for catchment area analysis. Here is a video comparing different cities in the UK and a brief explanation on how to read the dataviz.
"30 minute drive time analysis from major UK cities visualised as 3d coral geometry. 
The thickness of artery is proportional to the number of networks connected to it indicating busier routes. The falloff in height is linked to the proximity to the centre. 
Corals aren’t normalised in scale as the purpose of this is visualising the form and pattern the networks create. Animation is a boomerang motion scaling from 0 to 30 min and back again. Congestion/traffic not accounted for."

click at the bottom of the video to watch it in full screen and high definition

Yep, there are some obvious parallels here with Time Geography and in particular with the representation of space-time prisms. The static version of the space-time trees gives a sharper visualization of the data.

The space-time tree, or 3d coral geometry as Craig said.

and the inverted original dataviz, "the drive time web"

Thursday, May 3, 2018

Cities in Brazil: A Law and Economics Research Agenda

Just a few days ago, Edward Glaeser presented at the Harvard Law School Brazilian Association Legal Symposium (video below). Glaeser talked about his recent research and some of the questions it raises towards a research agenda on various challenges faced by cities in Brazil but also in other countries from the developing world. This is a self-recommendation post, I haven't watched the full video yet. Hat tip Bruno Bodart.

ps. curious fact mentioned in the video. Glaeser's PhD thesis advisor at Chicago was the Brazilian economist José Scheinkman.