From Twitter to Traffic Predictor: Next-Day Morning Traffic Prediction Using Social Media Data

The research addresses limitations in forecasting early-morning traffic congestion using conventional methods. We proposed mining Twitter data to understand how evening and overnight behavioral patterns influence next-morning traffic conditions.

Key findings indicate that the earlier people rest as indicated from tweets, the more congested roads will be in the next morning. The study also revealed that major events reflected in tweet sentiment and nighttime social media activity correlate with congestion patterns. The predictive framework uses tweeting profiles collected by 5 a.m. to forecast morning commute congestion.

Tested on Pittsburgh freeway networks, the approach considerably outperforms methods without Twitter message features and provides predictive advantages particularly for road segments upstream of bottlenecks with high day-to-day congestion variation.