HowlinWaits

Wait time analyzer for Howlin' Rays

Currently a WIP

Getting Started

First, you will to authenticate with 3rd party APIs. This project requires a secret key from the DarkSky API and OAuth 1.0 information for the Twitter API. For Twitter, you will only need minimal read-only privelages. Next, follow the SecretTemplate.py to populate HowlinWaits/Config/Secret.py.

You can run HowlinWaits with GPU training in Docker or with CPU training natively.

Host (cpu)

cd HowlinWaits
pip3 install .

If Tensorflow 2.1 is not found, check that you are running Python 3.6.x 64 bit and that your pip is up to date. To run using Tensorflow CPU just enter:

python3 WaitAnalayzer.py

Docker (gpu)

Install Docker and Nvidia-Docker.

Run the following in the project's root directory to build the docker image:

docker build -t howlin .

To run, simply execute ./run.sh

To do

Data Gathering:

• Fetch tweets
• Parse wait times from tweets
• Insert data into sqlite3 DB

Data Analysis:

• Determine analysis method
• Implement analysis method

Publish

• Create website displaying best times