Applying data science to increase market transparency

Integrating approaches from machine learning and operations research in resource allocation processes
- Examples include applying NLP tools for identifying breakthrough patents to evaluating product differentiation across companies, leveraging API or web-scraped data to improve opportunity identification across an investment universe, for example Stripe, Shopify or Square data to understand relative strength and funding needs for small businesses, or growth in retail investor interest across different companies over time

The liquidity/volatility dual and related effects

ML approaches to quantifying liquidity premia following market dislocations
- How much are people willing to pay to enter/exit an asset with urgency?
- Is this premium forecastable?
- Can we use it to improve returns on an institutional scale?
- Can the segregation of retail order flow (and tighter spreads available to pooled retail order) create appealing niche opportunities?
Impact of changes in market structure on liquidity and asset returns
- How are changes in the relative abundance of different market participants and their behavior affecting market functioning?
  - Examples include connections between increasing herding behavior and return kurtosis and Gabaix and Koojien’s inelastic market hypothesis

Exchanges are an appealing model given their focus on capturing a risk-less spread, shifting the operator’s key concern from identifying the direction of price moves to promoting trade volume
- Can this model be used to create more efficient markets in niche areas such as small business funding, which currently suffer from poor transparency and mismatches in participant scale?
- Can lessons from niche exchanges, for example prediction markets or sports betting exchanges be applied to creating other socially beneficial markets?