Data Analyst · Ontario, Canada

Dhyey
Jani

I build Power BI dashboards, Python models, and data stories that help businesses move from raw numbers to clear decisions. Focused on business intelligence, machine learning, and analytics that are useful from day one.

1
Case Study
4
Power BI Dashboards
25+
SQL Queries
5
Python & ML Projects
Power BI · Sales Dashboard Live
$3.1M
↑ 14%
Revenue
8.3K
↑ 9%
Orders
94%
↑ 3%
On-Time
Python · ML Model Accuracy: 87%
model_eval.py
# Predictive model summary
accuracy_score(y_test, y_pred)
# → 0.871
Top feature: recency 0.34
Precision: 0.84 Recall: 0.78
# Model evaluation complete
print("Training accuracy: 0.89")
Feature importance calculated
# Cross-validation scores
np.mean(cv_scores)
# → 0.856
# Predictive model summary
accuracy_score(y_test, y_pred)
# → 0.871
Top feature: recency 0.34
Precision: 0.84 Recall: 0.78
# Model evaluation complete
print("Training accuracy: 0.89")
Feature importance calculated
# Cross-validation scores
np.mean(cv_scores)
# → 0.856
About

Turning data into
decisions that matter.

I'm Dhyey Jani, a data analyst based in Ontario, Canada, building a portfolio centred on business intelligence, dashboarding, machine learning, automation, and data storytelling. My focus is on work that is analytically sound and immediately useful to business stakeholders.

My Power BI dashboards are designed around clarity and interactivity, giving decision-makers filterable views of performance without needing to open a spreadsheet. My Python and notebook projects follow the same standard: structured, reproducible analysis with findings that can be acted on.

I'm actively building toward a full-time data analyst or BI analyst role where I can contribute meaningful analytical work from day one.

BI & Visualisation
Tableau Power BI DAX Power Query Excel Dashboard Design KPI Reporting
Programming
Python Pandas NumPy Scikit-learn Matplotlib
Analysis
Data Cleaning EDA Machine Learning Trend Analysis Business Insights
Tools
GitHub Jupyter Notebook VS Code
My Work

Projects & Case Study

Explore my portfolio of data analytics work across business intelligence, SQL analysis, and machine learning. Each project demonstrates practical problem-solving using highly realistic data that mirrors complex business environments.

Power BI Dashboards

Interactive Business Intelligence

Below is a collection of interactive dashboards I've built using Power BI. Each project highlights a unique data story, whether it's tracking trends, uncovering insights, or visualizing performance.

These dashboards were created by connecting, transforming, and modeling data from various sources, then designing visuals that make complex information easy to explore.

Feel free to click through, apply filters, and interact with the data to dive deeper into each project.

Sales Trends
$2.8M
Revenue
↑ 11%
vs Prior
Power BI Dashboards
Sales Trend Report

Track sales performance across regions, products, and time periods with interactive filters and trend analysis.

Revenue breakdown by region and product category
Time-based trend analysis with period-over-period comparison
Drill-through capabilities for detailed transaction views
Power BIDAX Sales Analytics
Financial Performance
$1.9M
Revenue
94.2%
Budget
Power BI Business Performance
Financial Trend Report

Monitor financial metrics, revenue trends, and budget analysis to support business planning and period reviews.

Revenue and expense tracking with budget variance analysis
KPI summary tiles for executive-level performance overview
Time-based comparisons with flexible period slicers
Power BIDAX Financial Analytics
Supply Chain Report
$578K
Revenue
46,099
Units
Power BI Supply Chain
Supply Chain Report

Monitor supply chain operations, inventory levels, and supplier performance across multiple operational views.

Revenue, stock levels, and cost tracking in KPI summary cards
Product and supplier analysis with SKU-level drill-down
Multi-page dashboard supporting overview and detailed analysis
Power BIDAX Supply Chain Analytics
Sales & Revenue
$3.1M
Revenue
↑ 14%
Growth
Power BI Business Performance
Sales and Revenue Analysis Report

Track overall sales performance and revenue trends to support business reviews and strategic planning decisions.

Revenue and sales volume tracked against prior period with variance
KPI summary tiles give executives an at-a-glance performance view
Trend lines and period slicers allow flexible time-based comparison
Power BIDAX Revenue Analysis
Case Study

Dashboard Deep Dive

Supply Chain Analytics Dashboard
Domain
Supply Chain Analytics
Type
Interactive Power BI Dashboard
Focus
Revenue, product movement, supplier analysis, stock visibility
Tools
Power BI, DAX, Power Query
Filters
SKU, Supplier name
Pages
Overview, Product, Supplier
The Problem

Supply chain data is often spread across separate views of revenue, inventory, product activity, and supplier performance, making it difficult to review overall operational health in one place. This report was designed to bring those areas together in a clearer dashboard structure that supports both summary-level review and more focused analysis.

How are revenue, stock levels, product movement, and supplier performance trending across the supply chain?
Project Note

Built as a Power BI practice project inspired by a public supply chain dashboard pattern, with my own analysis refinements and presentation decisions.

Dashboard Design

The report was built in Power BI as a multi-page dashboard with Overview, Product, and Supplier sections. KPI cards highlight core business measures including Total Revenue ($578K), Total Product Sold (46,099), Total Costs (58K), Stock Level (4,777), and Avg Profit Margin % (86.07).

Additional visuals break performance down by product type, SKU, shipping carriers, transportation modes, and supplier-level analysis. SKU and Supplier name slicers allow users to narrow the dashboard to specific operational segments without manually filtering source data.

Key Insights the Dashboard Surfaces
01

The overview page gives a quick snapshot of commercial and operational performance through KPI cards and category-level revenue views.

02

The product page brings together order quantity, stock level, lead time, and profit margin visuals to support product-level monitoring.

03

The supplier page focuses on supplier analysis, transportation costs, stock contribution, and SKU-level product relationships for vendor review.

Value Delivered

This dashboard turns multiple supply chain views into a more connected analytical report that is easier to navigate and interpret. Instead of reviewing separate tables or exports, users can move between overview, product, and supplier perspectives while keeping the same filtering context.

$578K
Total revenue surfaced on the overview page
46,099
Total product sold visible as a top-line KPI
4,777
Stock level tracked directly in the dashboard
SQL Projects

Data Querying & Analysis

A comprehensive SQL analysis project using a sales database with four relational tables: sales transactions, products, geographical regions, and salesperson data. The queries below demonstrate data retrieval, filtering, calculated columns, aggregation, and conditional logic used to extract business insights from the database.

Explore sales performance across products, regions, and salespeople through interactive visualizations built from the SQL database.

SQL Database Schema showing relationships between geo, sales, people, and products tables

Database structure showing four relational tables with primary and foreign key constraints. Sales data connects to geography, products, and salesperson tables.

A collection of SQL queries demonstrating data retrieval, filtering, calculated columns, aggregation, and conditional logic.

Select everything from sales table
SELECT * FROM sales;
Show just a few columns from sales table
SELECT SaleDate, Amount, Customers FROM sales;
SELECT Amount, Customers, GeoID FROM sales;
Adding a calculated column with SQL
SELECT SaleDate, Amount, Boxes, Amount / boxes FROM sales;
Naming a field with AS in SQL
SELECT SaleDate, Amount, Boxes, Amount / boxes AS 'Amount per box' FROM sales;
Using WHERE Clause in SQL
SELECT * FROM sales
WHERE amount > 10000;
Showing sales data where amount is greater than 10,000 by descending order
SELECT * FROM sales
WHERE amount > 10000
ORDER BY amount DESC;
Showing sales data where geography is g1 by product ID & descending order of amounts
SELECT * FROM sales
WHERE geoid='g1'
ORDER BY PID, Amount DESC;
Working with dates in SQL
SELECT * FROM sales
WHERE amount > 10000 AND SaleDate >= '2022-01-01';
Using year() function to select all data in a specific year
SELECT SaleDate, Amount FROM sales
WHERE amount > 10000 AND year(SaleDate) = 2022
ORDER BY amount DESC;
BETWEEN condition in SQL with < & > operators
SELECT * FROM sales
WHERE boxes > 0 AND boxes <= 50;
Using the between operator in SQL
SELECT * FROM sales
WHERE boxes BETWEEN 0 AND 50;
Using weekday() function in SQL
SELECT SaleDate, Amount, Boxes, weekday(SaleDate) AS 'Day of week'
FROM sales
WHERE weekday(SaleDate) = 4;
Working with People table
SELECT * FROM people;
OR operator in SQL
SELECT * FROM people
WHERE team = 'Delish' OR team = 'Jucies';
IN operator in SQL
SELECT * FROM people
WHERE team IN ('Delish','Jucies');
LIKE operator in SQL
SELECT * FROM people
WHERE salesperson LIKE 'B%';

SELECT * FROM people
WHERE salesperson LIKE '%B%';
Using CASE to create branching logic in SQL
SELECT SaleDate, Amount,
        CASE WHEN amount < 1000 THEN 'Under 1k'
             WHEN amount < 5000 THEN 'Under 5k'
             WHEN amount < 10000 THEN 'Under 10k'
             ELSE '10k or more'
        END AS 'Amount category'
FROM sales;
GROUP BY in SQL
SELECT team, count(*) FROM people
GROUP BY team;
Python & Jupyter

Python & Notebook Projects

These projects cover machine learning, process automation, customer analytics, and social media analysis. All built in Python and published on GitHub. Each represents a structured analytical workflow from problem framing through to interpretation, readable directly in a browser.

What these projects show

Python-based analytical work lives in GitHub repositories and Jupyter notebooks, readable as both runnable code and narrative reports. Each project below links directly to the file or report, viewable without any setup by recruiters or hiring managers.

🤖Machine learning models with evaluation metrics and documented interpretation
📓Jupyter notebooks combining live code, visualisations, and written commentary
⚙️Automation scripts for operational and workflow improvement tasks
📊Customer and social analytics with actionable segment-level findings
Final Report.pdf
# Predictive model evaluation
model.fit(X_train, y_train)
score = model.score(X_test, y_test)
# → Model ready for inference
Report: documented ✓
Machine Learning
Machine Learning · Python
ML Predictive Modelling

A structured predictive modelling project applying machine learning techniques to a real dataset. Covers model selection, training, evaluation, and documented interpretation of results in a final analytical report.

End-to-end ML pipeline from data preparation through model evaluation
Model performance assessed using standard classification or regression metrics
Findings documented in a structured, shareable analytical report
PythonScikit-learn PandasPredictive Modelling
mnist_classifier.py
# MNIST digit classification
model.fit(X_train, y_train)
Test accuracy: 98.3%
# 10 classes: digits 0–9
Confusion matrix: logged ✓
Machine Learning
Machine Learning · Python
ML MNIST Database

A machine learning project using the MNIST handwritten digit dataset to build, train, and evaluate classification models. Exploring model architectures and benchmarking performance on a well-known image recognition problem.

Classification model trained and evaluated on the standard MNIST dataset
Experiments with model architecture and hyperparameter choices
Results documented with confusion matrix analysis and accuracy metrics
PythonScikit-learn NumPyClassification
OCCU Automation.ipynb Python 3
# Automated workflow script
for record in data:
  process(record)
  log_output(record)
→ Workflow complete. 0 errors.
Automation
Jupyter Notebook · Automation · Python
Ontario Community Credit Union Automation

A process automation project built for a credit union context. Python scripting and a notebook-based workflow replace a repetitive manual data handling task, reducing effort and improving consistency.

Scripted automation replaces a manual, error-prone operational process
Notebook structure makes the workflow transparent and auditable
Reproducible and rerunnable with updated input data
PythonJupyter PandasAutomation
segmentation_report.pdf
# Customer segmentation
KMeans(n_clusters=4).fit(rfm)
Segment A: High-value 28%
Segment B: At-risk 19%
Report: documented ✓
Customer Insights
Analytics · Python · Customer Insights
Customer Segmentation Analysis

A customer analytics project using clustering techniques to identify meaningful customer segments from transactional or behavioural data. Supporting targeted marketing and retention decisions.

Cluster analysis applied to group customers by shared behaviour patterns
Segment profiles described with actionable business implications
Findings published as a structured analytical report
PythonPandas Scikit-learnSegmentation
TrendStream Analysis.ipynb Python 3
# Engagement trend analysis
df.groupby('platform')['engagement']
  .agg(['mean', 'max'])
→ Peak engagement: Thursday 7pm
Social Analytics
Jupyter Notebook · Python · Social Media Analytics
TrendStream Social Media Analysis

A Jupyter notebook exploring patterns, trends, and engagement metrics across social or content-related data. Surfacing platform-level and time-based insights into what drives content performance.

Engagement metrics analysed by platform, content type, and time of day
Trend visualisations highlight peak performance windows and patterns
Notebook format combines code, charts, and written interpretation in one document
PythonJupyter PandasMatplotlib
Contact

Let's Connect

Open to opportunities.
Let's start a conversation.

I'm actively building my analytics portfolio and exploring roles that align with my skills. Whether you're hiring, want to collaborate on a project, share experiences, or offer suggestions and feedback. I'd be glad to hear from you through any of the channels below.

Resume & contact

If you have any questions, feedback, or just want to connect, feel free to reach out. I'm always open to learning, collaborating, and new opportunities.

Whether you're working on something in data, looking for someone to join your team, or simply want to share thoughts. I'd be happy to hear from you. Every connection is a chance to grow and improve, and I'm excited to keep learning as I go.