jpressnell / cloud-roles-database

Cloud Roles Database

Open Cloud Data

Motivation

I created this project after getting asked many questions about getting into the cloud industry, and rather than giving my own anecdotal experience too much, I was curious to gather data. I decided one of the easiest places to start would be with publicly available job description information. I started first by gathering data into a spreadsheet, and whilst that was a useful learning exercise, that process was painful and slow... so I automated it. In future, this project can be extended with additional information such as from surveys.

Getting Started

This project ingests open-source information, such as information hosted on job descriptions, loads it into a SQL (Postgres) database and creates a website with the information for analysis.

Currently, the project hosts no database, but runs a three step process:

1. Ingest (make run-ingest) - Starts the crawling for data (depends upon make db_migrate already being invoked to setup the DB).
2. Compilation (make run-query-compilation) - Compiles SQL queries into static JSON files used for rendering.
3. Rendering as HTML (make run-render) - This command renders the JSON files as a static site.

Roadmap

Data Ingest

• Salary & pay information
• Seniority breakdowns
• Soft skills
• Role overlap
• Specialties, e.g. Kmubernetes vs Serverless
• Update reference data based on Stack Overflow survey (e.g. languages and databases)
• Work experience
• Location
• Update information on languages and monitoring
• Certifications
• Niche cloud providers (digital ocean, etc)
• FAANG data

Insights

• Delta of the cloud roles to the main data set
• Most popular practices (not skills)
• Most popular:
  • databases
  • infra as code tools
  • programming language
  • cloud provider
• Should you learn one cloud provider or two (and the overlap)?
• Which roles in the cloud use Kubernetes?
• Which is the most popular certification?

Ingest

• Fix regex for Go / AI / Java / JavaScript (e.g. select only body?)
• Better logging (verbosity levels) and error state handling for ingest
• Make URL the ID of the jobs in database
• Add historical date and data support
• Add unit tests for various resumes to test the import logic

Compilation

• Create a file per file to render
• Simplify SQL to JSON export (e.g. direct in Postgres)

Rendering

• Add a role detail page (based on a YAML config)
• Provide meta information (ingest count, role count)