Predictive Modeling · Google Data Analytics Capstone

FC Barcelona Match Predictor

A logistic regression model trained on five seasons of FC Barcelona match data. Set the matchup conditions and the model returns win, draw, and loss probabilities along with an expected goals estimate.

247Matches analyzed
5Seasons (2019-2024)
6Model features
71%Test accuracy
Interactive Model

Predict a result

Adjust the inputs below. The model recalculates instantly using the same feature weights learned during training.

Top side = title rivals / Champions League level. Lower half = relegation-zone clubs.
10
15 = five straight wins. 0 = five straight losses.
Most likely outcome
Win
Model confidence
Win0%
Draw0%
Loss0%
Projected goals for: 0.0 Projected goals against: 0.0
Exploratory Analysis

What the data shows

Five seasons of results cleaned and structured for analysis. A few patterns stood out before any modeling began.

Result distribution

Across all 247 matches

Home vs away win rate

Camp Nou advantage is real and measurable

Win rate by opponent tier

Results compress sharply against top sides

Goals per match by season

Scored vs conceded, 2019-2024
Methodology

How the model works

A multinomial logistic regression predicts the probability of each outcome from six features. Weights below are scaled for readability and show each feature's relative pull toward a Barcelona win.

+0.84
Venue
Playing at Camp Nou is the single strongest positive signal.
+0.61
Recent form
Points won in the last five matches.
-0.92
Opponent tier
Facing a top side is the strongest negative signal.
+0.33
Rest advantage
Days of recovery relative to the opponent.
+0.47
Goal diff (season)
Rolling scoring margin entering the match.
+0.29
Head-to-head
Historical record against that specific club.
Collect & clean 247 matches Engineer 6 features Train / test split (80/20) Fit logistic regression Validate (71% accuracy) Deploy as live tool