Historical news for Symbol
- Thread starter nitro
- Start date
http://finance.yahoo.com/q/h?s=MSFT&t=2011-09-06T20:50:07-04:00&id=50866273
start collecting..you might also check some paid providers like IQ feed,esignal etc.
keep in mind,that you not going to get ALL the news. what you going to get is the news from participating provider. entirely possible for example that upgrade\downgrade from some small firm that could trigger the move is not going to be there.
start collecting..you might also check some paid providers like IQ feed,esignal etc.
keep in mind,that you not going to get ALL the news. what you going to get is the news from participating provider. entirely possible for example that upgrade\downgrade from some small firm that could trigger the move is not going to be there.
you can collect news data from yahoo for each ticker,each day by the end of the day
http://feeds.finance.yahoo.com/rss/2.0/headline?s=MSFT®ion=US&lang=en-US
however-for some time yahoo(like for a year or so) does not provide the source of the news(news provider,agency name), if you look or scrap the data from rss.
Hi Nitro,
From this and other posts I deduce that you are putting together a fairly impressive market database, suitable for use in bulding daily-frequency trading models. And that you are building this data-base cheaply, on a shoe-string, scraping data from various public sites.
Am I correct? If so how is it going? What is the overall plan, how extensive is the database going to be? Do you have an overall schema?
Also is this only for your private use, or will you make it available to others?
Well, I have several months of 1 minute bar (200 GB compressed) equity/ETF options data (so no, not HF) for the entire term structure of several symbols (in other words, the Mass Quote snapshot at one minute intervals along with highly accurate greeks and other things like higher moments that I compute). I have ideas for a model/system, and I need to gather simple things like earnings, corporate actions, dividends, news, etc, so that I can make sense of vola changes that I/model see. The _trading_ model is very sophisticated, but it mostly tries to quantify what traders do by eye. Although, it uses simple _options_ models for IV and greeks, where eventually I would like to use more sophisticated options models like Heston-Nandi to explain not just IV, but skew and kurtosis. Still, for this idea I don't think it is that important to use more sophisticated option model as input to the trading model.
It is rough going, but it is going. The database is perhaps too small, although I may be able to quintuple it. I am building these analysis/trading tools using Mircrosoft tools (C#, Sql Server, Sql Server Management studio, SSRS, SSAS, StreamInsight, etc etc etc), which are fantastic and you can download them for free (incredible and seriously MSFT is a God-send in this sense.) I am hoping for a C# version of Quantlib soon.
My "Schema" is very simple. I have come to the conclusion that avoiding joins as much as possible is critical. On the other hand, I don't have to go to the extreme that others go to here in compression and storing in binary (or HDF5) databases. I enjoy using the power of Microsoft SQL server management studio, and for me having few tables and very little joins, but still having the power of the SQL tools, is a sweet spot that works for me.