plot_timestamp

Attributes

model_rename_dict
TIME_BINS
predictions_csv

Functions

plot_score_over_time_bins(result_df, filename, score_col)	Visualize each forecaster's average performance over time bins (time before market close).
plot_time_gap_distribution(forecasts_df, filename[, ...])	Plot the empirical distribution of time gaps (time before market close) for each forecaster.
plot_time_gap_histogram(forecasts_df, filename[, ...])	Plot a histogram of time gaps for a specific forecaster (for debugging/exploration).

Module Contents

plot_timestamp.model_rename_dict

plot_timestamp.TIME_BINS

plot_timestamp.plot_score_over_time_bins(result_df: pandas.DataFrame, filename: str, score_col: str, forecasters: list[str] = None, time_bins: list = None)

Visualize each forecaster’s average performance over time bins (time before market close).

Args:

result_df: DataFrame with columns [‘forecaster’, ‘event_ticker’, ‘time_bin’, ‘weight’, score_col] filename: filename to save the plot score_col: column to plot (‘brier_score’ or ‘average_return’) forecasters: list of forecasters to plot. If None, plot all forecasters. time_bins: List of tuples (lower, upper, label) defining time bins. If None, uses default TIME_BINS

plot_timestamp.plot_time_gap_distribution(forecasts_df: pandas.DataFrame, filename: str, forecasters: list[str] = None, time_bins: list = None)

Plot the empirical distribution of time gaps (time before market close) for each forecaster.

Args:

forecasts_df: DataFrame with columns [‘forecaster’, ‘time_to_last’, ‘time_bin’] filename: filename to save the plot forecasters: list of forecasters to plot. If None, plot all forecasters. time_bins: List of tuples (lower, upper, label) defining time bins. If None, uses default TIME_BINS

plot_timestamp.plot_time_gap_histogram(forecasts_df: pandas.DataFrame, filename: str, forecaster: str = None, max_hours: float = 200)

Plot a histogram of time gaps for a specific forecaster (for debugging/exploration).

Args:

forecasts_df: DataFrame with ‘time_to_last’ column filename: filename to save the plot forecaster: specific forecaster to plot. If None, plots all forecasters combined. max_hours: maximum hours to show on x-axis (default: 200)

plot_timestamp.predictions_csv = 'slurm/predictions_10_01_to_09_01.csv'