DevopsDaily

 I decided to rip off the bandaid and use my Python skills to do some Data Science and retire R, as it really wasn't serving me to learn neural networks. The first thing I had to do was to figure out all the Python libraries and this is what I settled on:  a client library for Twitter, a current stock data library, and a stock technical analysis library. After that, I found a great open API for NLP sentiment analysis. The next library I needed was data storage and manipulation. I used a db library to store the data and, finally, Numpy and Panda to manipulate the data while experimenting.  This is what I built to solve the problem.  I took around 1,000 tweets daily with the search term Litecoin and did sentiment analysis on them to find out of the 42 tweets per hour how many of those are positive, negative, or neutral, and stored that in a MySQL database table.  Next, I took the stock data from Yahoo Finance and ran it through technical analysis software, looking...

I had the privilege of working among some of the most intelligent people I have ever known for several years while I was head of IT and DevOps at Skytree (The machine learning company) back in 2013-2016, way before the AI LLM hype that has taken over the world.  I learned from and worked with giants in the industry. Some of them went on to build their own companies, and some are legends in the field. While there, I got a real-world education in ML and what it could do at scales that hadn't been tried before.   In my off hours, I took Coursera courses on ML and the R Programing languages classes. After about the second year, I knew enough about our product and ML in general that I often stood in the booth, gave presentations about our product to passers-by at conferences, and sat in as a sales engineer at client meetings.   Now the problem: While at Skytree, I talked to a colleague about ML and inference. With enough candle stick technical analysis stock dat...