search

ProjectSidewalk / sidewalk-quality-analysis

An analysis of Project Sidewalk user quality based on interaction logs
5 stars 3 forks source link

To analyze where labels are placed on a street segment, I need the previously developed code #10

Open jonfroehlich opened 5 years ago

jonfroehlich commented 5 years ago

For some of our worker quality inference analyses, we will want to incorporate where users tended to place labels in terms of a street segment (e.g., distance from an intersection). This is the same analyses that we performed for the Sidewalk Computer Vision paper (link to comment). @tongning wrote this code.

image

@tongning can you point me to where you wrote this code so I can try and incorporate it into my Jupyter Notebook? Perhaps @galenweld may also have some insights...

galenweld commented 5 years ago

Anthony's code is super easy to use. You can download it from tongning/label-intersection-proximity and then you'll simply need to download the full_dc.geojson file or equivalent for whatever area you're interested in (Mikey was able to help me get this for Seattle and Newberg, it's just the raw data of the street grid used for computing the above).

Then, you'll need to install the dependencies - the only ones you'll not already have will probably be geojson and shapely

With this in place, simply call the function compute_proximity from Anthony's package and you'll be all set. More details are in his readme.

jonfroehlich commented 5 years ago

Nice! Thanks @galenweld. I'll look into incorporating this into my Jupyter Notebook!