Back in October, we hosted a Meetup in NYC where I discussed the potential of crowdsourced earnings estimates as the basis for new and interesting trading strategies. At the Meetup, I stepped through the different stages of Estimize's data in order to construct an algorithm that uses crowdsourced earnings as a trading signal. This algorithm uses the same logic as the algorithm released for EventVestor (Share Buyback Announcements) as well as the accompanying event study.
The notebook attached here was what I used to present to the Meetup and it goes through, in detail, what the data looks like and how the two averages (Estimize versus The Street) compare against each other. It also recreates an entire white paper, as presented by Estimize, in the Research Environment. So take a look and I promise you'll learn a lot.
Note: Aimed as more of an informative piece on Research, feel free to ask questions you might have about the process, executing code, and the transition between backtesting and the notebook. If you're curious about the data, send a message over to Estimize to find out more.