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/
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)