Beta v. 0.6

What's Changed

• Old main by @gmalbert in #9
• Revert "Old main" by @gmalbert in #10
• Update .gitignore, pyvenv.cfg, and add .f1.code-workspace file by @gmalbert in #11
• Update .gitignore by @gmalbert in #12
• Update .gitignore and fix tab functionality for MAE analysis. by @gmalbert in #13
• Update visualization components to use 'width' parameter for better l… by @gmalbert in #14
• Enhance podium and position displays with prediction error metrics an… by @gmalbert in #15
• Added features to deal with drivers changing teams by @gmalbert in #16
• Update readme with table of contents by @gmalbert in #17
• Add requirements.txt for streamlit cloud by @gmalbert in #18
• Remove .venv from repository by @gmalbert in #19
• Remove excess files by @gmalbert in #20
• Add documentation, remove files, and update readme by @gmalbert in #21
• Merge pull request #21 from gmalbert/main by @gmalbert in #22
• Changed to tab functions for top-level functions by @gmalbert in #23
• New logo with updates to the ReadMe by @gmalbert in #24
• Add function to dynamically configure dataframe height by @gmalbert in #25
• Update/dataframe height by @gmalbert in #26
• Git workflow tips by @gmalbert in #27
• Update to Readme and Roadmap, and data updates by @gmalbert in #28
• Update Next Race to show historical MAE by @gmalbert in #29
• Update with historical MAE in table by @gmalbert in #30
• Update/fix runtime by @gmalbert in #31
• Updated Predictive MAE calculations by @gmalbert in #32
• Update after Sao Paolo 2025 by @gmalbert in #33
• Update race analysis to add highlighted week and other cleanup by @gmalbert in #34
• Fix to row highlighting by @gmalbert in #35
• Update correlation matrix by @gmalbert in #36
• Removed unnecessary .copy functions and cleaned up Vegas pit constant by @gmalbert in #37
• Set model features to text files for easier importing, cleanup of com… by @gmalbert in #38
• Update to f1db 2025.21.1 by @gmalbert in #39
• Implement multi-model support and hyperparameter optimization by @gmalbert in #40
• Fixes to data models for Streamlit cloud by @gmalbert in #41
• Fix to XGBoost for Streamlit cloud by @gmalbert in #42
• streamlit cloud errors by @gmalbert in #43
• Addtl fixes for Cloud by @gmalbert in #44
• Server fix to avoid caching by @gmalbert in #45
• Add cache invalidation docs, deployment guide, and ignore CatBoost te… by @gmalbert in #46
• Fix to CACHE VERSION by @gmalbert in #47
• Minor fix to caching issue - WIP by @gmalbert in #48
• WIP - caching fix on cloud by @gmalbert in #49
• Added temporal leakage, reorganized tabs by @gmalbert in #50
• Fixed filter columns to get rid of extra filters from new fields by @gmalbert in #51
• Fix errant DNF filter and fix issue where qualifying data by @gmalbert in #52

Full Changelog: v0.4-beta...v0.6-beta

Beta v. 0.5

• Added predicted final position high and low (taking into account MAE)
• Added an input entry to modify the Early Stopping threshold
• Updated the MAE breakdown to be by winner, podium, top 10, and then the remainder

Full Changelog: https://github.com/gmalbert/f1Analysis/commits/v0.5-beta

Beta v. 0.4

• Added fastest pit stop to Next Race section
• Added check for High Cardinality
• Added early stopping details
• Added comparing different bin counts
• Modified high cardinality features to bins to enhance model
• Programmatically removed Jos Verstappen from possible matches from the drivers.json file because it was incorrectly matching to him rather than his son, Max Verstappen (which caused Max's data to not appear)

Beta v. 0.3

• Removed all commented out code for cleanliness
• Modified model to XGBoost.Booster to allow for early stopping rounds to limit overfitting
• Modified Feature Importances to Importances which measures how many times a feature went through the boosting rounds (to attempt to lower MAE and reduce overfitting)

Full Changelog: v0.2-beta...v0.3-beta

Beta v. 0.2

• Standardize Andrea Kimi Antonelli to Kimi Antonelli everywhere based on [f1db/f1db] Release v2025.15.0
• Removed SafetyCarStatus from data model which was causing data leakage
• Updated weather module to no longer re-pull data if it was already in the dataframe
• Fixed the issue with best_s1_sec, best_s2_sec, best_s3_sec, best_theory_lap_sec not being populated in dataframe

Beta v. 0.1

• Added feature list to README after Monte Carlo simulation
• Added approximately 40 features to data model
• Reduced MAE to 1.50
• Added Recursive Feature Elimination (RFE) and Boruta to streamline data model
• Updated DNF data model
• Introduced trending for both drivers and constructors to include 3 and 5 year trends as well as last 3 and 5 races to show recent improvements