60
Steady number of activities showing consistency.
4.68
Optimized pace demonstrates efficiency gains.
13.4
Endurance is increasing with the longest distance noted.
2.5
Expected runs per week reflect regularity.
These graphs show how musical features of songs listened to during runs relate to run performance. Each point represents a run, colored by distance (left) or pace (right). The arrows show how different musical features (e.g., tempo, energy) influence the run metrics. Longer arrows have more impact. For instance, high-tempo, acoustic, and instrumental music seems to correlate with longer distances and faster paces. However, very energetic songs might lead to early fatigue, potentially reducing overall pace.
While tuesdays show the lowest running frequency, they have the highest likelihood of achievements. This suggests tuesdays might be an optimal day for focused, high-quality runs. Used to go to office very monday, it, and thus avoided running sundays/mondays which lead to a nice recovery period.
The longest run was 12.34 km with a pace of 6.53 min/km.
Temperature at start of run of 23.140999°C and humidity of 46.933865%.
wind speed of 5.3150725 m/s.
Music features during the longest run: Acousticness: 0.29, Danceability: 0.47, Energy: 0.56, Instrumentalness: 0.84, Liveness: 0.23. Loudness: -7.90, Speechiness: 0.01, Tempo: 98.54, Time Signature: 11.00, Valence: 0.44.
The longest run was 4.41 km with a pace of 9.07 min/km.
Temperature at start of run of nan°C and humidity of nan%.
wind speed of nan m/s.
Two Kudos Prediction models were made: XGBoost and Random Forest. The models were trained on the Strava dataset and evaluated on the test set. The results are summarized in the table below:
The top 7 features influencing kudos count for the XGB model are visualized in the feature importance plots below.