niimpy.analysis.rhythms module
- niimpy.analysis.rhythms.compute_rhythms(df: DataFrame, timebin: str, cols: list, groupby_cols: list, period: int, freq: str, group: str | None = None) DataFrame[source]
Returns a pandas DataFrame containing rhythm computations for the input data based on specified frequency.
- Parameters:
- dfpandas.DataFrame
Input dataframe with a Timestamp index and ‘user’ column.
- timebinstr
Time bin for grouping, using pandas frequency string (e.g., ‘1h’, ‘2h’, ‘4h’, ‘6h’).
- colslist
List of columns to compute the count distribution for.
- groupby_colslist
Columns by which to group the data.
- periodint
Duration in hours for which rhythm is computed. E.g., 8 weeks would be 8*7*24 hours.
- freqstr
Sampling frequency, either ‘daily’ or ‘weekly’.
- Returns:
- rhythmspandas.DataFrame
Dataframe detailing call count distribution rhythms.
- Raises:
- ValueError:
If provided frequency is incorrect or a specified column isn’t in the dataframe.
- niimpy.analysis.rhythms.rhythm(df, period='4W', freq='1W', bin='1D', cols=[], groupby_cols=None)[source]
Compute rhythms from the input data.
Resample data to the given frequency.
Aggregate data to the specified period.
Compute the distribution over the frequency for the specified columns.
- Parameters:
- dfpandas.DataFrame
Input dataframe with a Timestamp index.
- periodstr
Period for which to compute the rhythm. Default is ‘M’ for monthly.
- freqstr
Frequency for the rhythm. Default is ‘W’ for weekly.
- binstr
Time bin for the rhythm computation. Defaults to ‘D’ for daily.
- colslist
List of columns to compute the count distribution for.
- groupby_colslist
Columns by which to group the data. Defaults to user and device
- Returns:
- rhythmspandas.DataFrame
Dataframe detailing rhythms.