Seems there are fewer and fewer arbitrage in the market… I have been trying to correlate various market data (prices of various indices and stock prices, volumes, order book, fundamental data) but the prediction is not really very accurate. (I trade both fundamental and quantitative)… I guess the reason being these data is widely available and any arbitrage is already captured by large hedge funds. Therefore recently I am exploring public data that is less readily available. For example, social media information and websit information. For example for e-commerce stocks, we can track down the number of SKU, number of brands, number of views of the website, identity of ads partners as some of the early predictor of their performance using data scrapping technology before their annual report announcement… and I can compare these parameters across several industry players for long-short strategy. I believe these kind of data will be the next big thing (and we should trade before this become popular and arbitrage is minimized again due to the use by the megasize hedge funds) I am wondering if you guys are exploring this area and leveraging other public data (may not be readily available) for your own trading? Let’s brainstorm.
What you've described can be generally classified as "social sentiment" and you're already a few years late on that one. That said, the implementation of such a model is much less precise than say a pure arbitrage where the values are concretely known. With regards to social sentiment, the big challenges are going to be filtering and weighting. This is not as straight forward. As such, if you come up with a better way to filter and weight that data, you could be ahead, but I can assure you there are much better capitalized firms working on this very problem. Can you do it as good or better before they suck all opportunity from the market? That part is up to you. Good luck!