October 19, 2020

159 words 1 min read

vaexio/dash-120million-taxi-app

vaexio/dash-120million-taxi-app

Explore 120 million taxi trips in real time with Dash and Vaex

repo name vaexio/dash-120million-taxi-app
repo link https://github.com/vaexio/dash-120million-taxi-app
homepage
language Jupyter Notebook
size (curr.) 536 kB
stars (curr.) 67
created 2020-06-01
license MIT License

dash-120million-taxi-app

Explore 120 million taxi trips in real time with Dash and Vaex

Available at https://dash.vaex.io/

img

Running this app

Clone the repo

$ git clone https://github.com/vaexio/dash-120million-taxi-app

Run in debug mode:

$ python app.py

Make sure the data is downloaded from s3:

$ python getdata.py

(You can ignore the warning WARNING:ThreadPoolExecutor-1_31:urllib3.connectionpool:Connection pool is full, discarding connection: vaex.s3.amazonaws.com)

Run in production mode (make sure the data is downloaded if you stream from s3):

$ VAEX_NUM_THREADS=8 gunicorn -w 16 app:server -b 0.0.0.0:8050

Settings

Change settings in the dash app

$ export TAXI_PATH=/data/taxi/yellow_taxi_2012_zones.hdf5  # change the default s3 file
$ export VAEX_NUM_THREADS=16     # change the number of threads per process/worker
$ export DASH_CACHE_TIMEOUT=240  # increase cache timeout to 4 minutes
$ export DASH_CACHE_TIMEOUT=-1  # disable cache (useful for benchmarking)
comments powered by Disqus