TimeCopilot Forecaster (Anomaly Detection)¶

Import libraries¶

In [1]:

%load_ext autoreload
%autoreload 2

%load_ext autoreload %autoreload 2

In [2]:

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)

In [3]:

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

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

Out[3]:

	unique_id	ds	y
0	Taylor Swift	2022-07-01	21499
1	Taylor Swift	2022-07-02	16349
2	Taylor Swift	2022-07-03	15042
3	Taylor Swift	2022-07-04	14358
4	Taylor Swift	2022-07-05	18332
...	...	...	...
6577	Selena Gomez	2025-06-27	8348
6578	Selena Gomez	2025-06-28	8670
6579	Selena Gomez	2025-06-29	9167
6580	Selena Gomez	2025-06-30	7902
6581	Selena Gomez	2025-07-01	7405

6582 rows × 3 columns

Import the models you want to use¶

In [4]:

from timecopilot.models.stats import SeasonalNaive, Theta
from timecopilot.models.foundation.chronos import Chronos
from timecopilot.models.foundation.arima import Arima
from timecopilot.models.foundation.flowstate import FlowState

from timecopilot.models.stats import SeasonalNaive, Theta from timecopilot.models.foundation.chronos import Chronos from timecopilot.models.foundation.arima import Arima from timecopilot.models.foundation.flowstate import FlowState

 See https://github.com/google-research/timesfm/blob/master/README.md for updated APIs.

Create a TimeCopilotForecaster¶

In [16]:

tcf = TimeCopilotForecaster(
    models=[
        #Chronos(repo_id="amazon/chronos-bolt-mini"),
        FlowState(),
        #SeasonalNaive(),
        #Theta(),
    ]
)

tcf = TimeCopilotForecaster( models=[ #Chronos(repo_id="amazon/chronos-bolt-mini"), FlowState(), #SeasonalNaive(), #Theta(), ] )

Plot the data¶

In [17]:

tcf.plot(df)

tcf.plot(df)

Out[17]:

In [23]:

fcst_df = tcf.forecast(df, h=21, level=[0, 20, 40, 60, 80] )

fcst_df = tcf.forecast(df, h=21, level=[0, 20, 40, 60, 80] )

INFO:p-17745:t-8116313152:modeling_flowstate.py:__init__:Number of encoder parameters: 7885.8240000000005k
INFO:p-17745:t-8116313152:modeling_flowstate.py:__init__:Number of dencoder parameters: 1181.952k (14.99%)
100%|██████████| 1/1 [00:01<00:00,  1.85s/it]

In [28]:

tcf.plot(df, fcst_df, level=[80], max_insample_length=42)

tcf.plot(df, fcst_df, level=[80], max_insample_length=42)

Out[28]:

Detect anomalies¶

You can optionally specify the following parameters:

• freq: The frequency of your data (e.g., 'D' for daily, 'M' for monthly). If you skip it, it will be inferred by default.
• h: The forecast horizon, which is the number of periods to predict during cross validation (you can skip it, and it will use an inferred seasonality)

In [ ]:

anomalies_df = tcf.detect_anomalies(df=df, h=7, level=99)

anomalies_df = tcf.detect_anomalies(df=df, h=7, level=99)

In [31]:

anomalies_df

anomalies_df

Out[31]:

	unique_id	ds	cutoff	y	Chronos	Chronos-lo-99	Chronos-hi-99	Chronos-anomaly	FlowState	FlowState-lo-99	FlowState-hi-99	FlowState-anomaly	SeasonalNaive	SeasonalNaive-lo-99	SeasonalNaive-hi-99	SeasonalNaive-anomaly	Theta	Theta-lo-99	Theta-hi-99	Theta-anomaly
0	Beyoncé	2022-07-06	2022-07-05	11571	11145.848633	-26105.343962	48397.041228	False	11127.576172	-25824.447434	48079.599778	False	11633.0	-30482.516525	53748.516525	False	10975.327148	-31700.535951	53651.190248	False
1	Beyoncé	2022-07-07	2022-07-05	11046	11386.040039	-25865.152556	48637.232634	False	10919.440430	-26032.583176	47871.464035	False	11633.0	-30482.516525	53748.516525	False	10399.726562	-32276.136537	53075.589662	False
2	Beyoncé	2022-07-08	2022-07-05	10874	11603.713867	-25647.478728	48854.906462	False	10822.110352	-26129.913254	47774.133957	False	11633.0	-30482.516525	53748.516525	False	9824.126953	-32851.736146	52499.990052	False
3	Beyoncé	2022-07-09	2022-07-05	10471	11858.917969	-25392.274626	49110.110563	False	10851.082031	-26100.941575	47803.105637	False	11633.0	-30482.516525	53748.516525	False	9248.526367	-33427.336732	51924.389466	False
4	Beyoncé	2022-07-10	2022-07-05	10721	12076.591797	-25174.600798	49327.784392	False	10776.775391	-26175.248215	47728.798996	False	11633.0	-30482.516525	53748.516525	False	8672.926758	-34002.936341	51348.789857	False
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
6547	Taylor Swift	2025-06-27	2025-06-24	17837	21935.486328	-78538.233969	122409.206625	False	23186.416016	-74337.494379	120710.326411	False	21965.0	-108196.537131	152126.537131	False	21373.949219	-111072.688801	153820.587238	False
6548	Taylor Swift	2025-06-28	2025-06-24	17536	21935.486328	-78538.233969	122409.206625	False	22668.628906	-74855.281489	120192.539301	False	21965.0	-108196.537131	152126.537131	False	21367.833984	-111078.804035	153814.472004	False
6549	Taylor Swift	2025-06-29	2025-06-24	17960	21935.486328	-78538.233969	122409.206625	False	22706.308594	-74817.601801	120230.218989	False	21965.0	-108196.537131	152126.537131	False	21361.716797	-111084.921223	153808.354817	False
6550	Taylor Swift	2025-06-30	2025-06-24	15490	22124.191406	-78349.528891	122597.911703	False	22523.058594	-75000.851801	120046.968989	False	21965.0	-108196.537131	152126.537131	False	21355.601562	-111091.036457	153802.239582	False
6551	Taylor Swift	2025-07-01	2025-06-24	15127	22124.191406	-78349.528891	122597.911703	False	21546.013672	-75977.896723	119069.924067	False	21965.0	-108196.537131	152126.537131	False	21349.484375	-111097.153645	153796.122395	False

6552 rows × 20 columns

Plot results¶

In [32]:

tcf.plot(df, anomalies_df)

tcf.plot(df, anomalies_df)

Out[32]: