Optimal Skills

This project helps you identify the best set of skills to learn for maximum job opportunities and better salary outcomes. The goal is to combine two ideas: skill demand and skill pay. You will use job listings data to find which skills appear most often, which skills are linked with higher salaries, and which combinations give the best overall value. By the end, you will have a ranked list of “optimal skills” and a clear learning roadmap based on evidence.

Project Goal

Answer questions like:

Which skills are most demanded for entry-level analytics roles?
Which skills are more common in higher-paying listings?
What skill combination gives the best balance of demand and pay?

Step 1: Build your dataset

Collect 40 to 120 job listings for roles like Data Analyst, Business Analyst, and Reporting Analyst. Save these fields:

job_id
title
company
location
experience_min, experience_max
salary_min, salary_max (if available)
salary_text (raw)
date_posted (optional)
description_text

If only some listings contain salary, that is fine. You will use salary analysis only on the listings that include it.

Step 2: Clean salary and experience

Standardise salary into one unit (example: annual salary). Create a single salary value using the midpoint of salary_min and salary_max. Create a single experience value using the midpoint of experience_min and experience_max. Remove or flag listings with missing salary so they do not break salary comparisons.

Step 3: Create a skill dictionary and extract skill flags

Create a list of skills and keywords. Then generate binary columns like:

has_python, has_sql, has_excel, has_powerbi, has_tableau
has_statistics, has_data_cleaning, has_dashboard
has_cloud, has_etl, has_communication

Keep your matching approach transparent and simple. The aim is to demonstrate analysis logic, not build a perfect NLP model.

Step 4: Calculate skill demand score

For each skill, calculate:

demand_count: number of job posts that mention it
demand_share: percentage of total job posts that mention it

Convert this into a demand score (for example normalise between 0 and 1).

Step 5: Calculate skill pay score

Using only job posts with salary:

calculate average salary for jobs that mention the skill
calculate average salary for jobs that do not mention the skill
compute salary uplift for the skill (difference or percentage uplift)

Convert this into a pay score (normalise between 0 and 1).

Step 6: Build an optimal skill score

Create a combined score:

optimal_score = 0.6 × demand_score + 0.4 × pay_score

You can adjust weights based on your goal. If you want maximum interviews, demand weight can be higher. If you want higher salary, pay weight can be higher.