◦ Engagement KPIs Anomalies Alert Manager: Created a dashboard & Google Cloud Dataset for Zee5’s 5 key video engagement KPIs. Enables all Zee5 teams to monitor & receive alerts for video experience on the Alert Reporting dashboard. ◦ Upgrading and Automating of Engagement KPIs Alert Reporting: Designed and developed an automated emailing system for Engagement KPIs breakdown scenario. Developed mathematical equations which helped in identifying breakdown scenarios. Upgraded the traditional Excel dashboard where data was pushed manually to an automated looker dashboard that used to refresh every 12 hours. The work which earlier required 2 hours was now completed in 10 minutes. ◦ PlayStore Negative Reviews Analysis and Dashboard: Created a monitoring dashboard for Zee5 app negative reviews. Implemented sentiment analysis & clustering for topics. Assisted customer service & engineering, reducing Oct negative reviews by 20%. ◦ Device Metadata effects on Engagement Metrics: Conducted in-depth analysis on device impact on engagement KPIs, revealing crucial insights for Ads team targeting and business decisions. ◦ AI Meme Generator: Utilized Google Cloud Vertex AI for meme caption generation & curated Zee5 show imagery for the marketing studio team @ Zee Hackathon (Prompt Play). ◦ AppDiscovery: Worked on customer profiling and identifying for user research team to increase content discovery within new users. Customer profiling helped in identifying that 1.8 percent of paid users watch content less than a minute in a month.

◦ Bench-marking between PostgreSQL and GeoPandas Spatial Analysis: Worked on the project where benchmarked spatial analysis methods, specifically PostgreSQL and GeoPandas. The findings revealed that PostgreSQL outperformed GeoPandas by an impressive 60 times. These findings helped tech team to make major changes in the architecture.◦ Florida State Electric Stations Spatial Data Analysis and Dashboard: Designed and developed tableau dashboard for electric stations analysis in Florida state. Dashboard helped different clients in making key decisions such as ”opening new electric recharging station”

Project:- Social Listening> Created ML model to classify sentiments of tweets done on Piramal Twitter handle. After classifying tweets' sentiments, I worked on different fronts like NER(name entity recognition), Topic Modelling, and Matching the results received from NER to the database. The insights of this project were further usedby the marketing team for devising marketing strategies.> Results:- Achieved 88.6% accuracy in classifying sentiments of the tweets. It was found that only 7% of tweets were related to our customers and others were spam. Project:- Social Analytics> Preprocessed the raw data received from the data engineering team to draw out valuable insights. Data was from an employee's review website. The data was very cluttered.> Results:- Was able to draw out engagement rate, location, vintage, and reviews of the users from raw data. Was able to plot the likes and dislikes of employees on India's Map statewise. Project:- Anomaly Detection> Work was assigned to Identify the anomalies in the Piramal loan branches. Tried many algorithms for anomaly detection in that 14 input data.>Result:- Identified around 1.1% anomalies. Project:- Increasing Piramal App Penetration>Here task was given to me to identify the customers who are still using old methods to obtain documents when they can straight away download them from the Piramal App.>Results:-Was successful in that. Identified around 28% of customers still use old methods, and shared the data with the marketing team so that they can run a campaign on that. Project :- Chatter Classification> Created and tested various ML models for classifying tweets(tagged with 'Piramal' keywords) in different categories such as Spam, Grievance, Query, and Feedback. >Results:- Came up with a two-stage model, where we concatenated one SVM model with a word embedding model(GLOVE) to achieve the best results. An accuracy of 82% was achieved in classifying an unbalanced dataset.

Played an important role in finding eco-friendly products for enlistment on their platform and also dealing with the various suppliers of that product throughout the world. Worked on identifying customer-friendly features on the official website and tried to implement those changes. I used Balsamiq for the same.