Getting started

tsod is library for timeseries data. The format of a timeseries is always a Series and in some cases with a DatetimeIndex

1. Get data in the form of a a Series (see Data formats below)

2. Select one or more detectors e.g. RangeDetector or ConstantValueDetector

3. Define parameters (e.g. min/max, max rate of change) or…

4. Fit parameters based on normal data, i.e. without outliers

5. Detect outliers in any dataset

Example

>>> import pandas as pd
>>> from tsod import RangeDetector
>>> rd = RangeDetector(max_value=2.0)
>>> data = pd.Series([0.0, 1.0, 3.0]) # 3.0 is out of range i.e. an anomaly
>>> anom = rd.detect(data)
>>> anom
  0    False
  1    False
  2     True
  dtype: bool
>>> data[anom] # get anomalous data
2    3.0
dtype: float64
>>> data[~anom] # get normal data
0    0.0
1    1.0
dtype: float64
>>>

Saving and loading

# save a configured detector
cd = CombinedDetector([ConstantValueDetector(), RangeDetector()])
cd.fit(normal_data)
cd.save("detector.joblib")

# ... and then later load it from disk
my_detector = tsod.load("detector.joblib")
my_detector.detect(some_data)

Data formats

Converting data to a Series

import pandas as pd
df = pd.read_csv("mydata.csv", parse_dates=True, index_col=0)
my_series = df['water_level']

from mikeio import Dfs0
dfs = Dfs0('simple.dfs0')
df = dfs.to_dataframe()
my_series_2 = df['rainfall']