hey, here’s the things I’m looking at, what else should I be looking at?

I’ve been working on expanding a monetary policy research pipeline beyond simple hawkish versus dovish sentiment classification.

Instead of just tagging speeches as positive or negative, we’re now running a suite of specialized models that label specific dimensions inside Federal Reserve text, including:

• Stance (hawkish vs dovish)
• Certainty versus uncertainty
• Inflation relevancy
• Housing market relevancy
• Economic activity
• Money supply
• Foreign sector references
• Claim projection or forward-looking intensity

The idea is to move beyond “how does the Fed sound?” toward “what specific economic topics are driving the communication?” and to quantify topic relevancy in real time.

As a financial analyst, I’m curious what else people think is worth extracting from unstructured Fed data.

For example:

• Should I be modeling conditionality structure (if inflation persists… then…)?
• Measuring disagreement or dispersion across FOMC members?
• Tracking regime shifts in language before policy pivots?
• Extracting implicit reaction functions from repeated phrase structures?
• Linking topic emphasis to cross-asset volatility (rates, FX, equities)?
• Detecting narrative persistence versus abrupt topic rotation?

If you work with macro, rates, or systematic strategies, what signals have you found valuable from Fed speeches, minutes, or press conferences that go beyond simple sentiment scoring?