Matt Makai - Python web dev & Twilio Developer Evangelist.
There are several areas in which I need to improve to be a truly self-reliant full-stack software engineer. My full-stack focus is on web development, both traditional desktop and tablet browsers as well as constrained smartphone mobile browsers. This post describes those areas and what I'm doing through the next year to improve those deficiencies.
The technical areas I need to improve in are:
- Web design
- Server monitoring
- Data analysis & visualization
I also want to get better at content creation, marketing, and working with end users to understand their needs better, but I won't cover those less technical topics in this post.
Towards this end I will (finally) purchase a copy of Adobe Illustrator CS6, work on tutorials, and pick up tips from my web design friends. Retrofitting ProofDriven with new icons will likely be my first conversion project so I have a goal to work towards.
Lately I have started using Heroku to deploy many of my side projects, but I still do a lot of DevOps for clients with Linux, Apache, nginx, mod_wsgi, Green Unicorn, MySQL, and PostgreSQL.
I am ashamed to admit I have very little insight into how well the servers I'm using are holding up under load. This situation really needs to be corrected by using a combination of Graphite & statsd, Raven & Sentry (I do already use these in many cases), New Relic, and some quick and dirty custom code with Python's cProfile and psutil.
Monitoring is a complicated subject and every application has a different server resource utilization profile. Once I collect the metrics, my next area for improvement will be important.
There's something about working with Python that makes data analysis appealing. Maybe it's the easy-to-use built-in data structures or the abundance of analysis libraries such as pandas, NumPy, and SciPy. Whatever the reason, I plan to study O'Reilly's excellent Data Analysis with Python book to learn more, particularly about Pandas.
On the visualization side, d3.js is fantastic. I've already been using it to create simple visualizations such as prgrmrlvlup and some other internal Excella applications for business development. I'm a huge fan of using CSS to style the visualizations and the enter(), exit(), and data joins model. Overall d3.js will provide the foundation for visualizing the results of interesting data sets processed on the server side with Python. I'll write blog posts and add many d3.js-related entries in my learning log over the next several months.