Hi, I'm Sowmya Pallempati
A Passionate Data enthusiast with a strong affinity for exploring complex datasets, uncovering hidden patterns, and leveraging actionable insights to drive business success.
About
I am a graduate student in Business Analytics and Project Management at the University of Connecticut School of Business. With a solid foundation in data analysis, data modeling, and data visualization I possess the necessary skills to handle complex datasets and extract valuable insights.
During my tenure at Larsen and Toubro Infotech, I served as a Data analyst and Data engineer for two years. This experience allowed me to work extensively in handling large datasets, performing tasks such as data cleaning, manipulation, analysis, and implementing ETL data processes, optimizing data processing time. Notably, I contributed to the development of loan application platforms by creating interactive visualizations using Power BI to monitor approval statuses, resulting in enhanced customer experiences and a reduction in loan approval time.
- Languages: Python, SQL, Java, HTML, CSS
- Tools & Interfaces: Microsoft Excel, Power BI, Tableau, SAS, DBeaver, Oracle, MYSQL, JMP, Git, JIRA, MS-Office
- Analytical Skills: Data Analysis, Data Visualization, Regression modelling, Statistical Modelling, Clustering, Neural networks, Data Mining, Time Series Forecasting, Text Mining, Survival Analysis
- Frameworks: Pandas, NumPy, Matplotlib, Seaborn, pylab, Scikit-learn, TensorFlow, Keras.
I have demonstrated proficiency in diverse analytical techniques throughout various projects related to Healthcare, Finance/Insurance, and time series analysis and forecasting. These include exploratory data analysis (EDA), regression modeling, statistical modeling, clustering, neural networks, data mining and time series forecasting. My hands-on experience working with various tools has allowed me to effectively analyze data and present insights to stakeholders.
I am actively seeking Full Time opportunities in roles related to Business Analyst, Data Analyst, or Data science. I am particularly interested in joining a company that emphasizes teamwork, and collaboration, and prioritizes continuous learning and development. I am eager to contribute my skills and grow professionally within an environment that fosters both personal and career growth.
Experience
- Analyzed 17,580 log files to identify patterns causing display failures, leading to a 32% reduction in ad disruptions and averting potential revenue loss.
- Led end-to-end ETL and data preprocessing tasks with SQL, PostgreSQL, and Python resulting in a 40% improvement in data quality laying a robust foundation for subsequent analyses.
- Created excel VLOOKUPs and performed segmentation analysis by market types, enhancing operational efficiency by 15% through streamlined data analysis.
- Developed Power BI dashboards for real-time monitoring, reducing customer-reported issues by 28% through swift detection of playlist pattern inconsistencies for the content management team.
- Tools: Power BI, MySQL, Excel, Python
- Created dashboards using Tableau for identifying hotspot patterns and defects based on historic data of failed builds and consequences achieving a 20% increase in the defects detected before production.
- Implemented ETL pipelines and developed triggers, stored procedures, functions, and views in MySQL reducing manual data handling by 40%.
- Collaborated closely with cross-functional teams to identify data requirements and design dashboards. Co-ordinated with QA team to ensure successful sign-offs and resolve any SIT/UAT regression issues within GRC Framework.
- Tools: Tableau, MySQL, Excel, JIRA
- Developed a data-driven home-loan application platform by incorporating advanced data analytics techniques using SAS software to improve the loan approval process achieving a 30% increase in approval rates.
- Formulated performance statistics by creating 15+ measures using DAX expressions and built an interactive Power BI dashboard to monitor loan approval status, saving 100s of hours in analysis.
- Performed data manipulation and utilized analytical functions for trend identification and time-based analysis.
- Tools: PowerBI, SAS, Excel
Projects
A classification model predicting status of Insurance policies using Python
- Tools & Libraries: Python, Numpy, Sklearn, Collections, imblearn
- Developed and fine-tuned a Random Forest classification model in predicting policy cancellations to achieve an accuracy of 85%, leveraging advanced data analysis techniques such as data cleaning, EDA, feature selection, and hyperparameter tuning using grid search on a dataset with 700,000 data points and 16 features.
- Conducted model interpretability using permutation importance to identify the top 5 features that have the most significant impact on predicting policy cancellations, resulting in better decision-making and increased operational efficiency
An End-to-End data engineering pipeline for analysing uber data.
- Tools & Libraries:Mage-ai, Python, BigQuery, Looker studio, Lucidcharts
- Developed an optimized Entity-Relationship Diagram (ERD) for Uber's raw data, enhancing organization and relationships across 19 features and 1 million rows.
- Automated ETL processes by deploying code in Mage, streamlining operations. Utilized BigQuery on Google Cloud Platform to load 8 tables and implemented optimized SQL queries for efficient data retrieval.
- Crafted insightful Looker Studio dashboards, visually presenting key performance indicators (KPIs) including total revenue, average fare amount, and average trip distance.
A model Predicting heart strokes using JMP
- Tools & Libraries: Excel, JMP
- Performed Exploratory Data Analysis and Pre-processed raw data containing 5110 data points to make it fit for use with different models by drawing correlations and eliminating features with lesser influence.
- Explored and evaluated various models for predicting Heart Failure and obtained a prediction accuracy of 82.5% over the test data using Decision Tree Models.
Analysis to classify different trip types for Hilton Hotels using Customer Journey Analytics
- Tools & Interfaces: Customer Journey Analytics
- Identified patterns in Hilton customer data for classification of trips as business, leisure or bleisure to make personalized recommendations for each trip type resulting in an estimated increase in bookings by 8%.
- Evaluated trends in booking data and proposed impactful website modifications by analyzing traffic patterns and identifying gaps, leading to a 35% increase in customer satisfaction ratings
Forecasting Brazil's Electricity Demand using SAS Studio.
- Tools & Interfaces: SAS Studio, Excel, Proc SQL
- Developed and implemented time series analysis techniques to explore 23 years’ worth of electricity demand data in Brazil, identifying critical patterns, trends, and seasonality components.
- Utilized ARIMA modeling and stationarity tests to accurately forecast 2 years of electricity demand with an ARIMA(3,1,3) model yielding a lower MAPE of 1.2% and provided valuable recommendations for energy companies and policymakers in optimizing production and distribution schedules.
Analyzing and visualizing restaurant data for reducing Food Wastage and increasing revenue.
- Tools & Libraries: Excel, JMP, Power BI
- Collected 6 months of restaurant data and explored distributions and correlations among diverse revenue contributing factors.
- Built Power BI dashboard featuring key metrics for daily sales and food wastage. Redesigned the restaurants menu and processes, contributing to an estimated revenue increase of over 10% and an 18% reduction in wastage.
Analysis and creation of Interactive sales dashboard using Tableau.
- Tools & Interfaces: Tableau, Excel
- Created interactive Sales Dashboard for a Supermarket Chain using Tableau’s advanced features empowering stakeholders to visualize and analyze key performance indicators(KPIs) leading to data-driven decision-making
- Provided valuable insights to sales managers through comprehensive sales breakdown structure and summaries, resulting in a 15% increase in sales within the selected time frame.
A case study for optimizing cancer therapy outcomes.
- Tools & Interfaces: Excel, Python, Power BI
- Leveraged advanced EDA and methods like one-hot encoding, Lasso regression and ANOVA F-Test in predicting prematurely ending cancer therapies achieving an AUC score of 0.92 with CatBoost Classifier.
- Employed KNN clustering and provided segment-specific business recommendations saving Humana approx. $2,450,000 annually.
A collaborative filtering technique to recommend movies to users using python.
- Tools & Interfaces:Python, Pandas,SciPy, Seaborn, Sklearn
- Developed movie recommendation engine using item-based collaborative filtering and SVD techniques, achieving an accuracy of 85% based on user’s watch history and ratings.
- Created visualizations and incorporated the Pearson’s R correlations method, to recommend the top 10 movies with the highest correlation scores.
Certifications
.
Publications
Intelligent Traffic Management Using Big Data Analytics and IoT
Publication Journal: International Journal for Research in Applied Science and Engineering Technology (IJRASET)
Volume: Volume 9, Issue X, Oct 2021
Smart pillow: An intelligent Pillow to track and improve sleep
Publication Journal: The International Journal of Analytical and Experimental Modal analysis (IJAEMA)
Volume:Volume XII, Issue VIII, Aug 2020
Education
Hartford, USA
Degree: Master's in Business Analytics and Project Management
CGPA: 3.95/4.0
Hyderabad, India
Degree: Bachelor of Technology in Electronics and Instrumentation
CGPA: 3.90/4.0