NS Health & Population Analytics

01

Project Overview

Nova Scotia is experiencing significant demographic shifts — an aging population, changing socioeconomic conditions, and increasing pressure on its healthcare system. Despite this, there is no consolidated, data-driven view connecting population trends, socioeconomic indicators, and health outcomes across NS regions.

This project addresses that gap. Using publicly available data from Statistics Canada, the NS Open Data Portal, and CIHI, Team A builds an integrated Business Intelligence solution that surfaces regional health patterns, demographic trends, and forecasted healthcare demand — delivered as an interactive Power BI dashboard.

Key analytical questions

Q1How has Nova Scotia's population age distribution shifted over the past decade?

Q2Which regions have the highest rates of chronic disease and hospitalization?

Q3Is there a correlation between socioeconomic indicators (income, education) and health outcomes?

Q4How do Nova Scotia's health indicators compare to national averages?

Q5Which demographic segments are expected to drive future healthcare demand?

02

Project Progress

Week 1

Scoping & Setup

Project charter
GitHub repo
Data acquisition
Sprint plan

● In Progress

Week 2

Data Wrangling

SQL/Python cleaning
Power Query ETL
Data dictionary
Schema design

○ Upcoming

Week 3

Analysis & Modelling

Pandas EDA
Trend regression
Forecast CSV
PBI wireframe

○ Upcoming

Week 4

BI Dashboard

Power BI build
DAX measures
Slicers & drill
Internal review

○ Upcoming

Week 5

Final Delivery

Final report PDF
Repo tag v1.0
Peer evaluation
Reflection

○ Upcoming

03

Business Understanding

CRISP-DM Phase 1

Nova Scotia health authorities and policy planners currently lack a single integrated view of how demographic change relates to health outcomes at the regional level. This project provides that view to support better resource allocation and planning decisions.

Stakeholder	Interest
Government of Nova Scotia	Policy planning, healthcare resource allocation
NS Health Authorities	Operational planning, regional health monitoring
Data Analysts / Researchers	Data-driven insights and modelling
Instructor/Evaluator	Academic assessment
General Public	Deployment of production-level AI systems

In scope

Nova Scotia population demographics data
Health indicators (hospital usage, chronic diseases, aging
Data cleaning and preprocessing(Python/SQL)
Exploratory Data Analysis (EDA)
Dashboard development (Power BI)
Basic forecasting or trend analysis
Documentation and GitHub repository

Out of scope

Real-time hospital system intergration
Clinical diagnosis or medical predictions
Patient patient-level data
Deployment of production-level AI systems

04

Data Understanding

CRISP-DM Phase 2

Source	URL	Coverage
NS Open Data Portal	data.novascotia.ca	NS health authority datasets, demographics
Statistics Canada	www150.statcan.gc.ca	Population, socioeconomic, health survey data
CIHI Open Data	cihi.ca	Hospitalization, wait times, health workforce

Data types collected: population by age/gender, life expectancy by region, chronic disease prevalence rates, hospitalization statistics, income and education indicators.

📄 Data dictionary: /data/data_dictionary.md Week 2

05

Data Preparation

CRISP-DM Phase 3

All raw data is stored in /data/raw/ and must not be edited directly. Cleaned outputs are written to /data/cleaned/.

Script	Purpose	Status
scripts/01_cleaning.sql	SQL cleaning, joins, aggregations	Week 2
scripts/02_cleaning.py	Python/pandas cleaning pipeline	Week 2
scripts/03_power_query.md	Power Query M-code documentation	Week 2

Planned star schema

Dimension

DimRegion

Dimension

DimDemographic

→

Fact Table

FactHealthOutcome

←

Dimension

DimDate

Dimension

DimCondition

06

Modelling

CRISP-DM Phase 4

📓 Python notebook: /notebooks/analysis.ipynb Week 3

Step	Description
EDA	Age group distributions by region, correlation matrix of socioeconomic vs. health variables, seaborn heatmaps
Trend Analysis	Time-series trend fitting per age cohort and health region
Predictive Model	Linear or polynomial regression — 5–10 year population and healthcare demand forecast
Export	Forecast table exported as CSV for Power BI import

07

Dashboard

CRISP-DM Phase 5 & 6

📊 Power BI file: /powerbi/NS_Health_Analytics.pbix Week 4

Page	Key Visuals
Executive Summary	4 KPI cards: NS Population, Avg Life Expectancy, Top Health Region, YoY Change
Population Trends	Area chart by age group, population pyramid, NS choropleth map
Health Outcomes	Chronic disease bar chart, hospitalization line chart, income vs. health scatter
Demographic Deep Dive	Stacked age/gender bar chart, community-level indicator table
Forecast	Python-generated forecast overlay with confidence interval ribbon

08

Repository Structure

ns-health-population-analytics/
│
├── data/
│   ├── raw/                  # Original downloaded datasets — do not edit
│   ├── cleaned/              # Cleaned, transformed outputs
│   └── data_dictionary.md    # Field definitions and transformation notes
│
├── scripts/
│   ├── 01_cleaning.sql       # SQL cleaning and joins
│   ├── 02_cleaning.py        # Python/pandas cleaning pipeline
│   └── 03_power_query.md     # Power Query M-code documentation
│
├── notebooks/
│   └── analysis.ipynb        # EDA + predictive model + CSV exports
│
├── powerbi/
│   └── NS_Health_Analytics.pbix
│
├── docs/
│   ├── project_charter.docx
│   ├── internal_review_log.md
│   └── final_report.pdf
│
├── .gitignore
└── README.html
    

SQL / SQLite

Data cleaning, joins, aggregations

Python

pandas, scikit-learn, seaborn — EDA and modelling

Power Query

ETL transformations and reshaping

Power BI

Interactive dashboard and DAX measures

GitHub

Version control for all scripts and docs

Jupyter

Notebook for EDA and forecast model

09

Team

KT

Pham Thi Kim Thanh

Project Manager

Coordination, timeline management, communication

AM

Anne Mwangi

Data Lead

Data sourcing, structure, and schema design

CO

Chukwuma Okoro

SQL / ETL Analyst

Data cleaning and transformation pipelines

JA

Juliana Ayide

Reporting Lead

Power BI dashboard development and DAX

MO

Vacancy

QA / Documentation

Data validation, documentation, review log

10

How to Navigate This Repo

For reviewers & collaborators

1

Start with this README

Understand project context, scope, and structure before diving in.

2

Read the project charter

/docs/project_charter.docx — full scope, methodology, stakeholders, and milestones.

3

Check the data dictionary

/data/data_dictionary.md — every field defined with source, type, and transformation notes.

4

Follow the analysis notebook

/notebooks/analysis.ipynb — step-by-step EDA, trend analysis, and predictive model.

5

Open the Power BI dashboard

/powerbi/NS_Health_Analytics.pbix — requires Power BI Desktop.

To reproduce the analysis

1

Download raw data

From the three sources listed in Section 04. Place files in /data/raw/.

2

Run the cleaning scripts

scripts/01_cleaning.sql or scripts/02_cleaning.py to produce cleaned outputs.

3

Install Python dependencies

pip install pandas numpy scikit-learn seaborn matplotlib jupyter

4

Run the analysis notebook

Open notebooks/analysis.ipynb in Jupyter. Run all cells to reproduce EDA and export forecast CSV.

5

Load into Power BI

Import cleaned CSVs and forecast output into NS_Health_Analytics.pbix.

NS Health &Population Analytics

Project Overview

Project Progress

Business Understanding

Data Understanding

Data Preparation

Modelling

Dashboard

Repository Structure

Team

How to Navigate This Repo

NS Health &
Population Analytics