// data_analyst.portfolio — executive summary

Harish
Thiyagharajan Selvi

3+ years building analytics infrastructure that drives measurable decisions — from multi-million-row SQL pipelines to executive-ready Power BI dashboards. Currently at PointClickCare.

0^%

Query speed gain

0^%

Fulfillment time ↓

0^h

Weekly hrs saved

0^%

Report errors ↓

Impact metrics — % improvement by initiative

Source: PointClickCare (2025) · Hexaware Technologies (2021–2024) · Updated June 2025

// skills_proficiency

Tools & Technologies

SQL (T-SQL / PL-SQL)95%

Power BI + DAX90%

Python (pandas, sklearn)85%

Tableau65%

Azure (ADF, Synapse)60%

Snowflake / BigQuery60%

Competency radar

Full toolset

SQLT-SQLPL/SQL Power BIDAXPower Query PythonpandasNumPy scikit-learnmatplotlibseaborn SnowflakePostgreSQLMySQL BigQueryAzure ADFSynapse Analytics TableauLookerExcel VBAPower PivotGit JupyterJiraConfluence ServiceNowSharePointLucidchart

// work_experience

Professional Experience

PointClickCare

Data Analyst

Feb 2025 – Present
Ontario, Canada

40%

Query speed-up

33%

Fulfillment ↓

18%

Incidents ↓

Business units

Optimized complex SQL queries across multi-million-row operational datasets, reducing report execution time by ~40% and improving Power BI dashboard refresh reliability across 5 business units.
Designed Power BI dashboards (DAX, Power Query) tracking SLA compliance and workload distribution, directly informing 3 executive process redesign decisions.
Automated weekly data reconciliation between ServiceNow and SQL Server with Python (pandas), eliminating ~3 hours of manual effort per analyst per week.
Enforced data validation frameworks across 5 business units, reducing metric inconsistencies in executive reports by ~30% and analyst onboarding time by ~20%.
Produced proactive SLA risk analyses surfacing breach patterns before escalation, contributing to an 18% reduction in repeat incident volume within 60 days.

Hexaware Technologies

Data Analyst

Sep 2021 – Apr 2024
India

Automated/wk

Clients served

25%

Rework ↓

8M+

Max rows

Built and optimized SQL pipelines and Power BI reports across 8 client engagements in ITSM and process optimization, reducing reporting cycle time by ~2 days per sprint.
Performed EDA using Python (pandas, matplotlib, seaborn) on datasets ranging from 500K to 8M+ rows, uncovering process gaps that drove 3 confirmed improvement decisions.
Mapped manual reconciliation workflows and configured automation, eliminating 8 hours of weekly analyst effort across 3 client teams.
Established KPI definitions and data dictionaries across 6 engagements, reducing requirements rework by ~25% and accelerating dashboard build cycles.
Translated complex findings into executive-ready presentations for 6 client leadership teams, influencing process redesign decisions across 3 engagements.

// featured_projects

Projects

// project_01

Enterprise KPI Dashboard Suite

Built end-to-end SQL extraction pipelines from 5 fragmented data sources into unified Power BI dashboards covering KPI tracking, SLA performance, and workload trends. Replaced 6 manual spreadsheet reports. 100% adoption within 30 days of go-live.

100% adoption · 5 hrs saved/wk · 3 decisions influenced

SQLPower BIDAXPower Query

// project_02

Customer Churn Prediction Model

Random Forest binary classifier predicting telecom customer churn 30 days ahead. Feature-engineered 20+ variables including usage patterns, contract type, and billing behaviour. Deployed as an interactive Tableau dashboard with full code published on GitHub.

82% accuracy · 30-day early warning · GitHub published

Pythonscikit-learnRandom ForestTableau

// project_03

Toronto Housing Market EDA

Analysed 5 years of public housing data to surface price trend anomalies by neighbourhood and property type. Produced a 12-chart visual narrative covering seasonality, affordability index shifts, and supply-demand gaps across Toronto.

5 years of data · 12 charts · Neighbourhood-level insights

PythonpandasmatplotlibJupyter

// certifications_and_education