timoderbeste / timeseries_viz

timeseries_viz

This is a project for visualizing time series data.

For example, the canadian climate dataset shown below has the temperaure and precipitation data for the past 80 years.

You can visualize the dataset from three aspects: distribution, outlier, and correlation.

Distribution

time_viz = TimeseriesVisualizer(data_path, plt_style="ggplot")
time_viz.visualize_distribution(
  t_col="LOCAL_DATE",
  t_keys=["01-Jan", "02-Feb"],
  d_cols=["CALGARY", "TORONTO", "VANCOUVER"],
  t_key_rule="contains",
  d_col_pattern="MEAN_TEMPERATURE_%s",
  x_label="Year",
  y_label="Mean Temperature",
  x_ticks=np.arange(1940, 2021),
  save_path=os.path.join(data_dir, "distribution_plot.png"),
  show_plot=True,
)

Outlier

time_viz = TimeseriesVisualizer(data_path, plt_style="ggplot")
time_viz.visualize_outliers(
  t_col="LOCAL_DATE",
  t_keys=["01-Jan", "01-Feb", "01-Mar", "01-Apr", "01-May",
          "01-Jun", "01-Jul", "01-Aug", "01-Sep", "01-Oct", 
          "01-Nov", "01-Dec"],
  d_cols=["TORONTO", "VANCOUVER", "CALGARY"],
  t_key_rule="contains",
  d_col_pattern="MEAN_TEMPERATURE_%s",
  title="Outliers of Mean Temperature over the Past 80 Years",
  x_label="Month of the year",
  y_label="Mean Temperatures",
  x_ticks=["01-Jan", "01-Feb", "01-Mar", "01-Apr", "01-May",   
           "01-Jun", "01-Jul", "01-Aug", "01-Sep", "01-Oct", 
           "01-Nov", "01-Dec"],
  save_path=os.path.join(data_dir, "outliers_plot.png"),
)

Correlation

time_viz = TimeseriesVisualizer(data_path, plt_style="ggplot")
time_viz.visualize_correlations(
  t_col="LOCAL_DATE",
  t_keys=["01-Jan"],
  d_cols=["CALGARY", "EDMONTON", "HALIFAX", "MONCTON",
          "MONTREAL", "STJOHNS", "TORONTO", "VANCOUVER"],
  t_key_rule="contains",
  d_col_pattern1="MEAN_TEMPERATURE_%s",
  d_col_pattern2="TOTAL_PRECIPITATION_%s",
  x_label="Year",
  y_label="Pearson Correlation",
  x_ticks=np.arange(1940, 2021),
  save_path=os.path.join(data_dir, "correlation_plot.png")
)