Getting Started

Let's discover mockseries in less than 5 minutes 🚀

1. Install mockseries

pip install mockseries

2. Create a timeseries generator

A timeseries can be expressed as a combination of 3 components: trend, seasonality and noise.

The trend is the long term, average change of the timeseries, such as the increase in CO2 emissions.
Seasonality is a cyclic pattern in the timeseries, such as the impact of the day-night cycle.
Noise represents irregular and random changes of the timeseries.

from datetime import timedelta

from mockseries.trend import LinearTrend

from mockseries.seasonality import SinusoidalSeasonality

from mockseries.noise import RedNoise

trend = LinearTrend(coefficient=2, time_unit=timedelta(days=4), flat_base=100)

seasonality = SinusoidalSeasonality(amplitude=20, period=timedelta(days=7)) \

+ SinusoidalSeasonality(amplitude=4, period=timedelta(days=1))

noise = RedNoise(mean=0, std=3, correlation=0.5)

timeseries = trend + seasonality + noise

Here, 2 seasonalities with a different period are combined. It's as easy as an addition !
Many types of trends, seasonalities and noises are available.
Just combine them with the operators +, - and *.

3. Preview

Quickly plot your signals

# preview on minute, hour, day, month, year

seasonality.preview_week()

timeseries.preview_month()

##<style="width:100%; aspect-ratio:9/6;" src="">

       

4. Generate values

Sample your timeseries on a timeframe.

from datetime import datetime

from mockseries.utils import datetime_range

time_points = datetime_range(

granularity=timedelta(hours=1),

start_time=datetime(2021, 5, 31),

end_time=datetime(2021, 8, 30),

)

ts_values = timeseries.generate(time_points=time_points)

datetime_range helps you get the time points of a given granularity in a timeframe.
For instance:
timeframe of 1 hour, with a granularity of 1 minute: 60 points;
timeframe of 1 hour, with a granularity of 1 second: 3600 points.

5. Plot or write to csv

from mockseries.utils import plot_timeseries, write_csv

plot_timeseries(time_points, ts_values)

write_csv(time_points, ts_values, "hello_mockseries.csv")

You will get something like this:

Now it's your turn !
Go to the next page to learn how to combine signals.
Go directly to the API Reference to checkout the available types of trends, seasonalities and noises.