 Extracted user streaming service information, traffic data for 300,000+ subscribers focusing on 50 Music channels to identify listener’s preference for Channels and Music genres.  Performed data wrangling & feature engineering to identify top trending and popular channels, total unique users, total time spent by users in Python.  Developed predictive model to predict the popular songs for different genres across channels, discussed strategies to broadcast those songs, thus increased 3,000+ new listeners.

Worked on "Cabin Reliability Log Classification" project that decreased time and resources for manual corrective In-Cabin actions for Boeing 777-300 by building and automating the multi-class classification model of maintenance log files. Deployed the machine learning model in the production environment and successfully ran it on the test dataset. The precision and recall for all the categories was 90% and 92%. The model will be extended to automate the same maintenance process for other fleets. Worked on a project that predicted response time of Mileage Plus customer emails, that helped agents to categorize and prioritize important emails which lead to solving customer problems quickly. Worked on "Predicting Wi-Fi purchase take rate" project to build a predictive model to forecast the in-flight Wi-Fi take rate and proposed dynamic pricing strategy to increase the take rate by 10%. Ran 120 different models on DataRobot, and found an XG Boost tree model with take rate as target to be fastest and most accurate. The predicted Wi-Fi take rates model was created by using over 150 variables of customer, flight, ancillary and booking data. The model predicted take rate at a flight level, within 1.3% of actual take rate (or 2 customers) on average. Wrote SQL queries in Teradata SQL Assistant, AWS (Athena) for ad-hoc data pull requests from AWS (S3) and United's Enterprise Data Warehouse and analyzed Wi-Fi purchase patterns by finding outliers, missing values, perform variable transformation in AWS (Sagemaker)Built a prototype to solve the current bag tracking problems which notify passengers about their current bag status. This improved bag tracking experience would de-escalate the large quantity of potential customer problems because they would be aware of these cases sooner and can be proactive. Tools Used: Jupyter Notebook (Python Coding), Teradata SQL Assistant, Data Robot, AWS (S3, Athena, Sagemaker), MS-Excel

Assisted on the professors on her ongoing projects by performing data wrangling operations on huge datasets. Analyzed the datasets by performing exploratory data analysis, missing value imputations and variable transformationApplied Natural Language Processing, Sentiment Analysis techniques in Python to provide meaningful insights a textual data. Deployed classification models with the efficiency of 82% using various machine learning algorithms in Python. Developed interactive dashboards and visualizations worksheets in Tableau to understand important Key Performance Indicators (KPI)Project Description:The task was to analyze and classify all the reviews, ratings about the professors and courses posted by students on various websites and build a classification model that classifies all the reviews into positive and negative reviews. Also analyzed the average and top rating, scope of improvements of professors and courses of the summarized data in Python. Further used the summarized data to build interactive worksheets in Tableau to identify KPI for each semester and communicate them to advisors and professors Models used and results:To classify the reviews using Machine learning algorithms like Naïve- Bayes, Support Vector Machine, Random Forest, Decision Tree, Logistic Regression in Python.Top classifiersRandom Forest Classifier: 85% precision, 80% recallDecision Tree Classifier: 83% precision, 79% recallLogistic Regression: 78% precision, 85% recall Used K-fold cross-validation techniques in the current models to optimize the results. Tools Used: Jupyter Notebook (Python), Tableau, MS-Excel, MySQL Workbench

• Perform statistical analysis using Excel of the undergraduate and the graduate student’s dataset for the Global Engagement department collected from several heterogeneous sources• Manage and analyze the student database to inspect the student profiles based on their performances• Draft financial budget reports using student database to provide financial aid to students in an effective manner

Performed exploratory data analysis on 10+ million records with 100+ attributes, feature engineering & extracted the following numerical features from existing variables in Python.○ After applying pre-processing, feature engineering and feature selection, the following features were extracted:○ One month average call duration( (start time - end time) / total count) computed for inbound, outbound, International, and Roaming.○ Counts of inbound, outbound, international, roaming calls over six months.○ Account features such as account balance and account type(corporate or consumer).○ Plan type○ Billing features such as Bill Amount, Number of times average spending < Rs 300 in the past 6 months(binary).○ User features such as age, tenure, location, total revenue.•Deployed multiple machine learning classification modeling techniques to predict the likelihood of telecom customer churn and ended up using Gradient Boosting for prediction.Using the model predictions, the client retained 30% of the likely-to-be churn prepaid customers through churn management .Designed ad-hoc visualizations and impactful dashboards in Tableau to communicating actionable insights to clients.Tools used: Python (pandas, numpy, matplotlib, scikit-learn), Tableau

Conducted Joint Application Development (JAD) sessions with the end-users, SME’s and development team throughout SDLC to gather and analyze various requirements and data.Extracted, aggregated and manipulated huge data sets from multiple sources of various sales data of CardTrade.com. Performed data mining, statistical and regression techniques on the gathered data using Python packages and Excel. Examined & corrected data with issues like completeness, accuracy, redundancy using SSIS (ETL packages). Wrote complex SQL queries for easy data analysis among the different competitors and achieved efficiency by 10%. Created dashboards and visualizations in Tableau to analyze sales performances of various automobiles across India.

Maintained data as per requirement of management authority by the use of various aspects such as templates, graphs, charts, and pivot tables. Improved data by filtering out irrelevant information and prepared it for use in existing processes. Documented the technical and business requirements and worked with teams to reach the target on time. Developed interactive visualizations and dashboards as per the requirements from the stakeholders using Tableau