CITY Objective Perception

Data Visualization and Analysis of Street View Imagery in Relation to Income Equity in Manhattan

The first step of this project is to extract about 17,000 points along NYC street centerlines, and then to download these points' Street View Image(SVI) utilizing Google Static Street View API based on their location coordinates.

The second step is to utilize advanced Computer Vision algorithm(PSPNet) to make semantic segmentation for each street view image. This is to segment different categories in each image and give each category(tree, sky, car, building etc) a specific color.

The third step is to use python to calculate the final result of each category's ratio in each segmented image and join all those attributes to original street points with their longitude and latitude for further data visualization.

Individual Work by Yilin Wang

Explore Relation Between Objective Perception of Street View Image and Income Equity in Manhattan

After constructing the complete dataset, I imported the geojson file of 17000 street points with attributes of each categoty's ratio from segmentation into mapbox to make data visualization. To be specific, I choose green-view ratio and sky-view ratio for visualization. I extruded each circle based on their attributes on these two ratios, and made color gradients for each extrusion.

Then I visualized NYC median house income per block and made blue-color gradient as a basemap to compare with street view image objective perception.

Greenview Ratio of Each SVI

Click on each SVI for exact number of greenview ratio

→→ 0 to 1

Skyview Ratio of Each SVI

Click on each SVI for exact number of skyview ratio

→→ 0 to 1

Median Household Income per Block

$150,000 +
$100,000 - $150,000
$75,000 - $100,000
$50,000 - $75,000
$20,000 - $50,000
Less than $20,000

Explore Relation Between Objective Perception of Street View Image and Income Equity in Manhattan

In this page, the visualization further illustrates the correlation between green view, sky view, and median income across different blocks in New York City. Using a heatmap approach, it highlights street view images with green-view ratio and sky visibility and correlates these visual factors with the economic demographics of each area.

The intensity and gradient of colors represent the varying ratios, with green indicating green view and blue indicating sky view. As the color intensifies, it signifies a higher ratio, providing a clear visual cue to identify areas with more abundant greenery and sky-view elements.

Color Gradient Legend

Correlation Analysis of several Metrics of street view image derived from computer vision

Correlation Analysis