utils

Functions

calculate_time_to_last_submission(→ pandas.DataFrame)	Calculate the time gap from each prediction to the market close time.
assign_time_bins(→ pandas.DataFrame)	Assign each forecast to a time bin based on time to market close.

Module Contents

utils.calculate_time_to_last_submission(forecasts_df: pandas.DataFrame) → pandas.DataFrame

Calculate the time gap from each prediction to the market close time.

This function calculates how many hours before the market close each prediction was made. It uses the ‘close_time’ column (actual market close time from database) if available, otherwise falls back to approximating with the last submission’s snapshot_time.

Args:

forecasts_df: DataFrame with columns including [‘event_ticker’, ‘snapshot_time’]

Optionally includes ‘close_time’ for accurate close times

Returns:

DataFrame with an additional ‘time_to_last’ column (in hours) representing the time gap from this prediction to the market close time

utils.assign_time_bins(forecasts_df: pandas.DataFrame, time_bins: list) → pandas.DataFrame

Assign each forecast to a time bin based on time to market close.

Args:

forecasts_df: DataFrame with ‘time_to_last’ column (in hours) - time before market close time_bins: List of tuples (lower_bound, upper_bound, label) defining time bins.

Example: [(0, 6, “0-6h”), (6, 12, “6-12h”), …]

Returns:

DataFrame with an additional ‘time_bin’ column containing the bin label. Rows that don’t fit into any bin are filtered out.