STAX: Automated Time Series Forecasting Tool

Install

python setup.py install

Command Line Interface

python -m stax [-h] table column frequency output

Argument	Definition
table	The csv file with the time series data
column	Which column to forecast
frequency	Use daily or monthly data
output	Full directory for JSON results

Example Command

python -m stax ./data/airline-passengers.csv Passengers monthly airline-passengers-result.json

Example Results

You can call python -m stax.plot on a json result file to create matplotlib visualizations.

Example Output

{
    "meta": {
        "train": {
            "values": [
                112,
                ...
                407
            ]
        },
        "test": {
            "values": [
                362,
                405,
                ...
                432
            ]
        }
    },
    "models": {
        "ARIMA": {
            "model": "ARIMA",
            "test_predictions": [
                363.2909859213608,
                369.8852769813491,
                ...
                484.92738208204236
            ],
            "test_confidence_intervals": [
                [
                    323.8192989290492,
                    402.7626729136724
                ],
                ...
            ],
            "test_mean_absoloute_percent_error": 0.0949
        },
        "ExponentialSmoothing": {
            "model": "TripleExponentialSmoothing",
            "test_predictions": [
                343.11984637306654,
                ...
            ],
            "test_confidence_intervals": null,
            "test_mean_absoloute_percent_error": 0.0786
        }
    }
}

Webhook Server with Redis

1. Run the redis server
2. Spin up redis queue workers
3. Start the webhook server