Problem
The \TimeSeriesTensor._df2tensors()\ method in \7-TimeSeries/common/utils.py\ (and its copies in lesson subdirectories) uses the deprecated \DataFrame.as_matrix()\ method, which was removed in pandas 0.25.0+.
Current Behavior
Running the time series lessons with modern pandas versions raises:
\
AttributeError: 'DataFrame' object has no attribute 'as_matrix'
\\
Expected Behavior
The code should use the modern \ o_numpy()\ API, which has been the recommended approach since pandas 0.25.0.
Proposed Fix
Replace all occurrences of .as_matrix()\ with .to_numpy()\ in:
- \7-TimeSeries/common/utils.py\
- \7-TimeSeries/1-Introduction/working/common/utils.py\
- \7-TimeSeries/1-Introduction/solution/common/utils.py\
- \7-TimeSeries/2-ARIMA/working/common/utils.py\
- \7-TimeSeries/2-ARIMA/solution/common/utils.py\
Environment
- pandas >= 0.25.0
- Python 3.7+
Additional Context
This is a straightforward API modernization fix that ensures compatibility with current pandas versions while maintaining backward compatibility with the tutorial's educational goals.
Problem
The \TimeSeriesTensor._df2tensors()\ method in \7-TimeSeries/common/utils.py\ (and its copies in lesson subdirectories) uses the deprecated \DataFrame.as_matrix()\ method, which was removed in pandas 0.25.0+.
Current Behavior
Running the time series lessons with modern pandas versions raises:
\
AttributeError: 'DataFrame' object has no attribute 'as_matrix'
\\
Expected Behavior
The code should use the modern \ o_numpy()\ API, which has been the recommended approach since pandas 0.25.0.
Proposed Fix
Replace all occurrences of .as_matrix()\ with .to_numpy()\ in:
Environment
Additional Context
This is a straightforward API modernization fix that ensures compatibility with current pandas versions while maintaining backward compatibility with the tutorial's educational goals.