Recently in ABuddy Category

I put together an architecture diagram to show the components, the data stores, and how they interact.

After running version 1 of my ABuddy through a few test cycles, I figured there was enough potential to rework it into a more viable project.  I already had several improvements I wanted to make, and as I started thinking about it seriously, came up with several more.

In the 1st version everything happened at once: the quotes were harvested, the analysis was made and the static report file was generated for that day.  If anything failed along the way, the whole thing just horked.

Fixing all that was pretty much the motivation for reworking it into version 2.  I wanted these key features:

• Tolerance of network issues.
• Separation of data and presentation.
• Self-maintained reference database of analyst recommendation scores.
• Flexible rule engine.

First up, I separated everything out to distinct components.  Like so:

Klepto: The sole responsibility of the klepto is to acquire the raw quote data.

Duke: The duke is the keeper of all the analyst recommendations.

Professor: The professor's job is to perform the needed analysis and report the final scores.

I probably could have made the Professor into two components rather than one, but I didn't want to be breeding rabbits.  The analysis step is actually just a matter of collating all the gathered stuff into the report itself, so those two tasks fit nicely together.

So that's it.  ;)

Here's a screen shot of version 1 of my ABuddy.


You can see a few symbols being highlighted in here.  In this case, the green ones are good performers, but maybe not good to invest in right now (since they just jumped up). 

The red one might be a good investment, so I would investigate it.  But it also might not since it just dropped a chunk.  Hard to say from here. 

Actually, it's not so hard -- I can see that that's a bargain stock, because of the "OB" on the end.  That means it doesn't have enough financial stability to meet SEC requirements for something or other, or something like that.  Since I don't really know what I'm doing, I stay away from those.

I've tried playing with the stock market.  As a "busy person" it's very difficult for me to keep up with what's going on in the world.  The result: I lose.  I lose money, then I lose time trying to figure out how to not lose money, then I lose more time investigating dead-end stocks and funds, then I lose more time and money drinking and babbling about how many times I was almost rich and how hard it was growing up as a poor black boy in the South.


Finally, a few months ago, it dawned on me that I'm good at making computers do all my work for me.  I can spend hours getting the computer to do a 10 minute job for me, just so I don't have to do it.  So I decided to do a little experiment to see if I could get a computer to make worthy financial suggestions.

Thus was birthed ABuddy.

I slammed together version 1 as a proof of concept in an afternoon.  All it did was scrape quotes from Yahoo! pages, plunk them in a table and color code them if they dropped or rose by 5% or more.  I had it choose which symbols to look at by giving it a list of industries and having it pull the ones with the biggest market caps.

It was a complete shot in the dark, but for a first pass it did surprisingly well at narrowing down things for me to investigate.

I've now run 3 "test cycles" with this system in which I invested around $10k (of imaginary money) and tracked my success.  I learned a few things, and watched which patterns worked and which didn't. 

With those lessons under my belt, I've now started working on version 2.