Forecast Using the Chronos Family¶

Import libraries¶

In [1]:

import pandas as pd

from timecopilot import TimeCopilotForecaster

import pandas as pd from timecopilot import TimeCopilotForecaster

Load the dataset¶

The DataFrame must include at least the following columns:

• unique_id: Unique identifier for each time series (string)
• ds: Date column (datetime format)
• y: Target variable for forecasting (float format)

The pandas frequency will be inferred from the ds column, if not provided. If the seasonality is not provided, it will be inferred based on the frequency. If the horizon is not set, it will default to 2 times the inferred seasonality.

In [2]:

df = pd.read_csv(
    "https://timecopilot.s3.amazonaws.com/public/data/events_pageviews.csv",
    parse_dates=["ds"],
)
df.head()

df = pd.read_csv( "https://timecopilot.s3.amazonaws.com/public/data/events_pageviews.csv", parse_dates=["ds"], ) df.head()

Out[2]:

	unique_id	ds	y
0	Oktoberfest	2020-01-31	25376
1	Oktoberfest	2020-02-29	28470
2	Oktoberfest	2020-03-31	23816
3	Oktoberfest	2020-04-30	46186
4	Oktoberfest	2020-05-31	31213

Plot the data¶

In [3]:

TimeCopilotForecaster.plot(df)

TimeCopilotForecaster.plot(df)

Out[3]:

Import the Chronos class¶

In [ ]:

from timecopilot.models.foundation.chronos import Chronos
from timecopilot.models.stats import SeasonalNaive

from timecopilot.models.foundation.chronos import Chronos from timecopilot.models.stats import SeasonalNaive

Create a TimeCopilotForecaster¶

In [5]:

chronos_av_models = [
    "amazon/chronos-2",
    "amazon/chronos-bolt-tiny",
    "amazon/chronos-bolt-mini",
    "amazon/chronos-bolt-small",
    "amazon/chronos-bolt-base",
    "amazon/chronos-t5-tiny",
    "amazon/chronos-t5-mini",
    "amazon/chronos-t5-small",
    "amazon/chronos-t5-base",
    "amazon/chronos-t5-large",
]
chronos_models = [Chronos(repo_id=model, alias=model) for model in chronos_av_models]
models = chronos_models + [SeasonalNaive()]


tcf = TimeCopilotForecaster(models=models)

chronos_av_models = [ "amazon/chronos-2", "amazon/chronos-bolt-tiny", "amazon/chronos-bolt-mini", "amazon/chronos-bolt-small", "amazon/chronos-bolt-base", "amazon/chronos-t5-tiny", "amazon/chronos-t5-mini", "amazon/chronos-t5-small", "amazon/chronos-t5-base", "amazon/chronos-t5-large", ] chronos_models = [Chronos(repo_id=model, alias=model) for model in chronos_av_models] models = chronos_models + [SeasonalNaive()] tcf = TimeCopilotForecaster(models=models)

Generate forecast¶

You can optionally specify the following parameters:

• freq: The frequency of your data (e.g., 'D' for daily, 'M' for monthly)
• h: The forecast horizon, which is the number of periods to predict
• seasonality: The seasonal period of your data, which can be inferred if not provided

In [ ]:

level = [20, 40, 60, 80]
cv_df = tcf.cross_validation(df=df, h=12, level=level) 

level = [20, 40, 60, 80] cv_df = tcf.cross_validation(df=df, h=12, level=level)

In [7]:

tcf.plot(df, cv_df.drop(columns=["cutoff", "y"]), level=[80])

tcf.plot(df, cv_df.drop(columns=["cutoff", "y"]), level=[80])

Out[7]:

In [8]:

cv_df.head()

cv_df.head()

Out[8]:

	unique_id	ds	cutoff	y	amazon/chronos-2	amazon/chronos-2-lo-20	amazon/chronos-2-hi-20	amazon/chronos-2-lo-40	amazon/chronos-2-hi-40	amazon/chronos-2-lo-60	...	amazon/chronos-t5-large-hi-80	SeasonalNaive	SeasonalNaive-lo-20	SeasonalNaive-lo-40	SeasonalNaive-lo-60	SeasonalNaive-lo-80	SeasonalNaive-hi-20	SeasonalNaive-hi-40	SeasonalNaive-hi-60	SeasonalNaive-hi-80
0	Black Friday	2024-09-30	2024-08-31	2607	2080.0	1992.0	2176.0	1896.0	2288.0	1784.0	...	2832.483154	2160.0	518.105774	-1238.539673	-3294.386475	-6145.491211	3801.894287	5558.539551	7614.386719	10465.491211
1	Black Friday	2024-10-31	2024-08-31	2470	2688.0	2320.0	3328.0	2048.0	4352.0	1856.0	...	3999.947998	3746.0	2104.105713	347.460358	-1708.386597	-4559.491211	5387.894043	7144.539551	9200.386719	12051.491211
2	Black Friday	2024-11-30	2024-08-31	11058	18304.0	15936.0	20352.0	12352.0	22784.0	9152.0	...	21284.275391	17053.0	15411.105469	13654.459961	11598.613281	8747.508789	18694.894531	20451.539062	22507.386719	25358.490234
3	Black Friday	2024-12-31	2024-08-31	3548	3264.0	2736.0	4160.0	2352.0	5376.0	1992.0	...	3876.372070	5407.0	3765.105713	2008.460327	-47.386593	-2898.491211	7048.894043	8805.540039	10861.386719	13712.491211
4	Black Friday	2025-01-31	2024-08-31	1724	2008.0	1912.0	2080.0	1816.0	2192.0	1688.0	...	2504.032227	2656.0	1014.105774	-742.539673	-2798.386475	-5649.491211	4297.894043	6054.539551	8110.386719	10961.491211

5 rows × 103 columns

Evaluation¶

In [9]:

from functools import partial

from utilsforecast.evaluation import evaluate
from utilsforecast.losses import mase, scaled_crps

from functools import partial from utilsforecast.evaluation import evaluate from utilsforecast.losses import mase, scaled_crps

In [10]:

eval_df = evaluate(
    cv_df.drop(columns=["cutoff"]), 
    train_df=df.query("ds <= '2024-08-31'"), 
    metrics=[partial(mase, seasonality=12), scaled_crps],
    level=level,
)
eval_df.groupby("metric").mean(numeric_only=True).T.sort_values(by="scaled_crps").round(3)

eval_df = evaluate( cv_df.drop(columns=["cutoff"]), train_df=df.query("ds <= '2024-08-31'"), metrics=[partial(mase, seasonality=12), scaled_crps], level=level, ) eval_df.groupby("metric").mean(numeric_only=True).T.sort_values(by="scaled_crps").round(3)

Out[10]:

metric	mase	scaled_crps
amazon/chronos-t5-large	1.106	0.302
amazon/chronos-bolt-small	0.996	0.306
amazon/chronos-bolt-mini	1.282	0.331
amazon/chronos-bolt-tiny	1.152	0.336
amazon/chronos-bolt-base	1.335	0.350
amazon/chronos-t5-small	1.377	0.358
amazon/chronos-2	1.646	0.360
amazon/chronos-t5-base	1.419	0.379
amazon/chronos-t5-mini	1.426	0.428
amazon/chronos-t5-tiny	2.505	0.505
SeasonalNaive	1.576	0.521