New post! @timbryan000 uses text analysis to explore the NFL draft.
There's a lot of really interesting/useful stuff here, including a dataset of scouting reports and analysis of linguistic variation.
opensourcefootball.com/posts…
Open Source Football retweeted
Using publicly available data and open-source code, I wrote (!) about constructing NFL draft curves and explained why top picks aren't as valuable as we think they are (if they aren't being used on a quarterback)
On @Open_Source_FB: opensourcefootball.com/posts…
New post!
@JonasTrostle examines defensive efficiency against WR1s relative to other receivers.
Post: opensourcefootball.com/posts…
Source code: github.com/nflverse/open-sou…
New post! @richjand demonstrates the use of multilevel modeling to estimate NFL team ability on offense and defense
opensourcefootball.com/posts…
Open Source Football retweeted
NEW for @Open_Source_FB: using 🔥🔥torch🔥🔥 for R to classify coverages using NFL player tracking data with a computer vision approach
opensourcefootball.com/posts…
Open Source Football retweeted
❗ New post for @Open_Source_FB: I wrote down everything I know about making an NFL win probability model using xgboost in R:
opensourcefootball.com/posts…
New post from @Danderso13!
This Post is from an account that no longer exists. Learn more
New post! @bendominguez011 shows off how to use Python's sklearn to construct an NFL team strength model
opensourcefootball.com/posts…
New post! @jacklich10 looks at penalties in the NFL.
2 big things stand out:
--> Offenses have more control over penalties than defenses
--> Some types of penalties carry more predictive power than others
opensourcefootball.com/posts…
New post!
@adrian_stats homefield advantage in the NFL over time, finding it to be greatly diminished in the most recent two seasons
opensourcefootball.com/posts…
New post! From @jacklich10
Exploring best ways to factor in opponents when opponent-adjusting EPA
This figure here shows how the team tiers chart changes when taking opponents into account
opensourcefootball.com/posts…
New post!
@random_effects argues that more you look into the numbers, the better it looks for Aaron Rodgers
opensourcefootball.com/posts…
New post!
@awgymer uses the new roster functions in @nflfastR to look at trends in receiving by position over time
opensourcefootball.com/posts…
New post!
@richjand demonstrates the use of tidyModels to predict rush/pass decisions and looks at which teams have rushed or passed more than expected
opensourcefootball.com/posts…
New post!
@mikeyirene creates an Expected Field Goal metric, finding the all-powerful Justin Tucker No. 1 in field goals made above expected
opensourcefootball.com/posts…
New post!
@ThePuntRunts debuts an R package for puntalytics (!) and looks at punt return yardage allowed over expected
opensourcefootball.com/posts…
New post!
@reinhurdler takes a look at actual and expected fantasy points for receivers in 2019 using where players were targeted.
And look at this gorgeous table!
opensourcefootball.com/posts…
New post!
@benbbaldwin uses the new clock time column in nflfastR to re-examine defensive rest time, this time using actual time elapsed
opensourcefootball.com/posts…
