Fun With Enron and Glasshouse

Shows words which occur between 30 and 500 times within subject lines in the enron email database. Relationship lines are drawn from words to other words which occur in subjects with them, provided the parent word has no more than 40 children. This is just a fun exploration, demonstrating that the value of Glasshouse lies in discovery.

Category:

Science & Technology

Tags:

• cyberspace
• tropez
• meddle
• pink
• floyd
• glashouse
• enron
• data
• visualization
• green
• phosphor
• llc
• arkowitz
• ben
• lindquist
• graph
• afghanistan
• bush
• energy
• terrorism
• email
• virtual
• world

[arkowitz]

I expect to have to write some code which infers positive or negative impact from news items based on certain words...

[arkowitz]

@drmkkk Well we certainly want to relate using time; also there may be symbols in the news which match symbols in the forex, in some very specific cases. I would actually want to start with a 3d bar graph with time on one axis, type of forex/type of news on another, and height being the price or change in price in the case of forex, or positive or negative character of the news in the case of news items. Then we can look along the time axis and hopefully see patterns of correlation.

[drmkkk]

I can put together a CSV (importable to Excel) file with the news and it should be similar to this one done with Enron - just use the words as basis... however the more interesting part would be cross referencing this with data from the Forex. I also have Forex historical quotes for 30+ currencies (also CSV) but I am not exactly clear how the two may be cross-referenced (news vs prices)?!

So in essence we have 2 datasets (news + quotes) , if there is a way to relate them it should be worth it.

[arkowitz]

@drmkkk Got a dataset? This thing hooks up to Excel or mysql or other databases; wouldn't be hard to make a forex visualization... I guess the trick is parsing and structuring the news piece... which I could do...

[drmkkk]

pretty cool actually, do you have anything for forex data?

for ex. news vs price this would be good, I have forex expertise get in touch with me if interested