About Me
Data Analyst with hands-on experience in data validation, data quality checks, SQL, Python, Excel and Power BI. Experienced in analysing 46,000+ operational records and 10,000+ customer records, identifying missing value…
Data Analyst with hands-on experience in data validation, data quality checks, SQL, Python, Excel and Power BI. Experienced in analysing 46,000+ operational records and 10,000+ customer records, identifying missing values, duplicates, inconsistencies, trends and anomalies. Strong exposure to ETL, KPI monitoring, statistical analysis and business reporting, with financial dataset experience through a banking analytics project. BCA graduate with 8.16/10 CGPA.
Experience
Data Science Consultant
<p>Analysed 46,606 ITSM incident records across 25 attributes using SQL, Python, Excel and Power BI to identify data-quality issues, operational trends and SLA patterns.<br>
Performed data validation and preparation by identifying missing values, duplicate records, inconsistent fields and non-predictive attributes before analysis.<br>
Identified operational patterns linking repeated ticket reassignments with SLA breaches and reopened incidents with longer resolution times, supporting process improvement.<br>
Conducted monthly incident trend analysis and developed an ARIMA-based three-month forecast to support workload and resource planning.</p>
ITSM OPTIMISATION USING MACHINE LEARNING | Client Project
SQL | Python | Excel | Power BI
<p>Analysed 10,000+ customer records using SQL joins, filtering, aggregation and grouping to identify customer churn patterns.<br>
Performed data quality checks covering missing values, duplicates, formatting and consistency before dashboard development.<br>
Built a Power BI dashboard with 10+ KPIs, customer segmentation and trend analysis using DAX; identified 26.5% churn and approximately ₹139K estimated revenue loss.</p>
Python | SQL | XGBoost
<p>Analysed a financial banking dataset and engineered 20+ customer and campaign features to identify term-deposit subscription patterns.<br>
Evaluated classification models using ROC-AUC, Recall, F1-score and Confusion Matrix; optimised the classification threshold from 0.50 to 0.30, increasing identified true positives from 255 to 471.</p>