Aha well I think in able to export this data: for social network sentiment:https://psychsignal.com/

http://stocktwits.com/signals/heatmap#Technology or this in portable?

I am not really sure what a "trending ticker" is but I assume that it is a type of indicator that a stock will be moving either up or down based on external data sources.

Given this, do you want to be able to import data from Twitter in order to be able to predict the movement of certain stocks given trends in the social media, if so how can you get yours hands on this sort of data from the Twitter servers.

Regards
Steve

Worldwide or national trending signifies a large amount of volume tweeting about a ticker..either negative or positive...eg $JRJC. so yes steve exactly social media external data. That often signals volume> = price movement. And potential useful data source to measure psychological, behavioural effects ...and how these play out on stocks.

Ideally I'd love to create some formulae that could correlate : no of tweets+ sentiment of tweet ( measured through popular adjectives) with % of movement of stock x. I believe a simple linear chart could be so useful to start with. ( there must be this data out there already? )

The data source I posted above...pschosignal or whatever it's called already has all this info on an external data source but costs a lot of money to access.

I haven't really got anywhere with this due to lack of time..but interesting signal to accompany. Another signal.

Cheers alex

James from PsychSignal here. Our real time data is definitely pricey at this stage, but we do offer historical data for free and delayed options for less. Get in touch, happy to chat.

Also, IMHO I would think less about predicting price movement with social data / sentiment and more about predicting volatility. Check out this BigML post http://blog.bigml.com/2014/05/09/predicting-stock-swings/

Thankyou james. The more I researched the more I discovered...that social media would be useful more as a secondary signal. I'm yet to figure it out as am more focused on initial volume signal stage. But your : ' if I try to predict whether a stock simply went up or down, the model is barely better than flipping a coin; for whatever reason, it’s easier to predict a tech stock’s intraday volatility than its daily gain or loss. ' is insightful.

Due to my idea being more focused oncapturing large breakouts. Eg JRJC I don't know how useful social media would be apart from maybe measuring the length of the rally from breakout eg. volume of tweets is 1000% above average = how many days it will run.now that would be very interesting study. A bar chart maybe? Thereby using it asa sell signal.

Love to hear your thoughts thanks.

How deep does this social rabbit hole go? :)

Hi, I worked with some people years ago, we pulled news headlines from major online news sources and weighted them by sentiment. It was just an experiment to see if the sentiment data made any sense. Here are some sentiment charts i found from that experiment
https://www.dropbox.com/s/ykher94olvm382m/unnamed.jpg?dl=0
https://www.dropbox.com/s/fdygp2bt0tu80rf/unnamed.png?dl=0

We did not look into tweets since we had a more generic/macro interest in events of the world and how that might swing over time (talking about months here). Tweets are likely to spike up after a pump/dump, up to that point its just noise probably. Also twitter is full of bots spewing out tweets about this and that ticker all day long, what you want is the real tweets by real people and track the volume (or the unusual lack of volume...) of those.

Anyway, heres additional reading on sentiment stuff
http://www.battleofthequants.com/Research/sdarticle.pdf
http://misrc.umn.edu/workshops/2011/fall/OliviaSheng_Paper.pdf

Hi,
Yeah there is a free service that provide the tweet volume by stock ticker or by sector.
I guess that's quite knew so they are building up the service currently anyway you can already access to the data.
On this link you can fin the tweet volume updated live for each ticker of the nasdaq: http://stockrumorindicator.com/bystocks
This is an average and allow you to detect a stock that's really trending based on the tweet volume.
Hope that's helpfull.
Best,
Sebastien

Quantopian provides StockTwits free of charge https://www.quantopian.com/data/psychsignal/stocktwits. Fields in this dataset can easily be turned into factors and used in a pipline.

Yeah really good data on stocktwits. The point is that on http://stockrumorindicator.com/bystocks you can have unfiltered data. that's free of charge also for now. There you can really access to the tweet volume unfiltered and play with it. Just scrap the web page at bystocks level and it covers 8000 stocks on american market.