My Official business thread - Humble beginnings

[AAAgent]

Are you working on creating a platform/program that helps you identify trends in the markets so you can trade or invest off of them? I thought I read that somewhere at one point.

Also, if you're looking at selling a software as a service...I've worked for Bloomberg, Mergermarket, Debtwire, etc. which are all financial data and news products that function as a software as service type thing, let me know. I also follow a few people on youtube, that have built successful businesses by first documenting and talking about their industry. Essentially becoming an expert in the industry and then start selling their services using youtube to promote.

 

[synergy1]

Yes you have the right idea. I have experience with time series, machine learning, and data...but I don't want to go the strict algo route, and make esoteric forecasts if the yen/USD pair moves 3% in relation to the UK GDP. My system will actually be quite simple - determine trends using both data and market research ( to which your recommendations would be more than welcome). Next in the investable universe, I have 5 years ( soon to be 6) of GAAP financial data. The most you can get on google is 4, and you need to scrape it - mine is all in a relational database ready for whatever.

For example, suppose I want to get the best companies in regards to free cash flow yield per share, and return on invested capital, I could perform these queries on my own and rank these stocks to provide ideas.

Folks do what I do, but I think I can provide a unique edge to it someday. Its not a full time gig. And again, thanks in advance if you have any recommendations. I'll certainly take a look. If you have any interest in this, feel free to PM me . As always i'll post on these forums and am always open to swap ideas, strategies and what not.

cheers

 

[AAAgent]

There is this programmer that uses programming and linguistics to identify human emotional language on the web. He takes that languages and identify's trends based of consumer sentiment. He's built off this data for many years and although he can't use it to find great stock picks (which is what he built it for originally) he uses it to identify trends so he can capitalize off the trend as a whole as opposed to a stock as a whole. His idea was that the financial numbers are sometimes fudged on these public companies, governments instill policies that shift markets which create a lack of free market. Instead of basing his algo's off financials, it was better to base it off the language coming out directly from the market which is not manipulated.

 

[synergy1]

Yup exactly. I was planning on scraping transcripts of earnings calls and doing simple sentiment analysis to determine if a good company is doing well, or a bad company is doing poorly. This is not too difficult since there are many libraries in programs like R that do this. For example, scrape a transcript for a company like Tesla motors and perform poisson distribution analysis to determine which words are more prevelent. If I am finding deceptive words like non-gaap pop up, i'll be armed with one more piece of data for my thesis ( which will be based on fundamental analysis. In teslas example, i'll be looking at the cash burn rate, the current EPS, share dilution etc. Next i'll check the sentiment and see if its negative, and than perform more forensic analysis of their statements. As an example, I found tesla performed some creative accounting with their cashflow from operations - moreover the padded earnings by recognizing certain credits at an unusual time.

The goal is to combine value investing with data. Its all possible , but will take some time to get right. the programmer you mention is on the right path though!

 

[synergy1]

What a crazy road its been. We are officially live. Our platform takes payments, and fulfills orders ( well we do that manually). At this point I have touched most parts of the stack and am better at certain areas (backend) than others ( frontend).

Now comes the challenge of getting our platform onto googles index engine. I pounded out an SEO book and get the basic idea. We implemented a way to share content with an SEO friendly site map. Next up is doing this for the entire site and pinging google. It seems content is king, so now we will have to figure out what key words we want searches to come up with. perhaps some adwords campaigns and studies of the results.

Some folk learn programming pretty quickly, others take a while to get the hang of it. This thread was started at the beginning, and I am just now feeling comfortable with developing the entire stack. I've managed to deploy some cloud infrastructure, with a wsgi complaint django app behind an secure ( https) nginx server. Basically, front end, backend, quality and devops all rolled into one! The lesson, keep going even if it seems hopeless. you will learn.

Good luck to everyone out there!