timiczn/Ekiti-State-Unemployment-Survey-Analysis

Power BI analysis of an unemployment survey covering 921 respondents (post-cleaning) across Ekiti State, built to help an NGO's executive team make evidence-based decisions on training, placement, and outreach programs. # Ekiti-State-Unemployment-Survey-Analysis ### Contents - Project Overview - Data Source - Data Cleaning and Preparation - Data Model - Key Measures - Dashboard - Key Insights - Recommendations - Tools - Limitations ### Project Overview An NGO working on youth and graduate employment in Ekiti State collected survey responses from unemployed individuals to understand who they are, what they studied, which sectors they want to work in, and what skills they already have. The raw data arrived as a single flat Excel export with heavy free-text entry, duplicate submissions, and several multi-select fields stored as inconsistent semicolon-delimited text. This project turns that raw export into a clean, related data model with a working dashboard and an executive report, so the NGO can prioritize programs by actual respondent demand instead of assumption. click to interact with the live dashboard ### Data Source A single Excel sheet of raw survey submissions, one row per response, covering: full name, date of birth, age, age bracket, highest degree attained, school attended, year of graduation, phone number, email, gender, preferred sector(s), and skill set(s). The raw export contained **1,006 submissions**. After cleaning, **921** were retained for analysis (see below). ### Data Cleaning and Preparation | Step | Detail | |---|---| | Deduplication | Matched on Full Name + Date of Birth (more reliable than phone number, which had typo'd digits between repeat submissions). Kept the latest submission per person. Removed 51 duplicate rows. | | Invalid row filtering | Removed rows with implausible ages (0, 5, 125), missing gender, and a future graduation-year sentinel value. | | Text cleanup | Trimmed whitespace, fixed inconsistent casing (`Hnd` → `HND`, `Nce` → `NCE`), lowercased emails, stripped trailing delimiters. | | Age Bracket recalibration | Replaced vague labels ("Youth", "Young Adult", "Middle Age", "Older Adult") with explicit, unambiguous bins …