news.md
In klarsen1/MarketMatching: Market Matching and Causal Impact Inference

Package: MarketMatching

Version: 1.2.1

MarketMacthing version 1.2.1

Rebuilt based on R version 4.3.2
Removed ggplot settings that created warnings
Wraps as.Date() around the date variable to truncate
Improved error and warning messaging

MarketMacthing version 1.2.0

This is a major release for CRAN.
The biggest differences between this version and the previous version in CRAN are:
1. Prospective pseudo power analysis by testing the design against fake interventions.
2. Generate optimal test/control market splits.

MarketMacthing version 1.1.4

Changed MAPE calculation to not re-scale the residuals.
Minor bug fixes.
The fake lift testing function does not refit underlying BSTS model at every step on the curve.
Added the parameter suggest_market_splits to best_matches() to provide a market split recommendations.
You can now pass just 2 markets (no longer a requirement of 3+ markets).
Added more details on the numbers behind the distance calculations.
Not submitted to CRAN.

MarketMacthing version 1.1.4

Changed MAPE calculation to not re-scale the residuals.
Minor bug fixes.
test_fake_lift() does not refit underlying BSTS model at every step on the curve.
Added the parameter suggest_market_splits to best_matches() to provide a market split recommendations.
You can now pass just 2 markets (no longer a requirement of 3+ markets).
Added more details on the numbers behind the distance calculations.

MarketMacthing version 1.1.3

Added the function test_fake_lift() to calculate pseudo power curves. Fixed minor bugs.

MarketMacthing version 1.1.2

Fixed documentation errors.

Added additional input checks.

Refactored the code to handle errors with markets_to_be_matched not being in data. Check happens prior to the execution of the calc distances loop now.

Modified calculate_distances and best_matches to allow for parallel processing to be executed when the specifies markets for markets_to_be_matched.