Are Gas Prices Tied to Tweets? Uncovering Predictive Patterns in Public Discourse

Can digital gripes about gas prices reveal real-world trends before they happen? This post dives into how tweet frequency and sentiment may hint at shifts in fuel costs across regions.

Data Context and Methodology

• Required data: Tweet frequency, keywords ("gas price", "fill up"), price indices (AAA).
• Source: Twitter API, GasBuddy, AAA Weekly Reports.
• Cleaning: Group by region, filter spam, convert time zones.

Volume Correlation with Market Timing

Mentions peak 3-5 days before local price hikes — a possible early indicator?

Regional Sensitivity and Predictive Validity

Stronger correlation in urban counties. Less predictive in rural regions with fewer posts.

Forecasting Possibilities and Limitations

Could this model feed into a regional forecasting tool?

Link to data: Gas Price Twitter Correlation Dataset