Ajeet Kumar

1. HDAM (Household Data Adjustment Model):

• Developed and maintained a predictive model to estimate household demographics from TV viewing patterns.
• Developed and maintained a predictive model to estimate household demographics from TV viewing patterns.
• Leveraged Nielsen's panel data to refine and validate demographic predictions, implementing robust data quality checks and validation processes.

2. Subminute TV Audience Measurement Application:

• Pioneered a novel application to reverse-engineer tuning information at a sub-minute level from known demographics.
• Utilized advanced PySpark techniques to analyze high-volume television viewing data, significantly enhancing the granularity and accuracy of Nielsen's audience measurement capabilities.

3. Big Data Processing and Optimization:

• Designed and optimized ETL processes and big data pipelines in AWS for both HDAM and sub-minute applications, reducing processing time, and enabling more frequent updates to household demographic assignments and tuning predictions.
• Implemented efficient data processing strategies to handle the increased granularity of sub-minute level data without compromising system performance.

4. Data Analysis and Reporting:

• Conducted an in-depth analysis of TV viewing patterns and demographic predictions using Python and PySpark.
• Developed automated reporting and visualization tools to communicate insights from both HDAM and sub-minute data to stakeholders, enhancing decision-making processes.

5. Cross-functional collaboration and continuous improvement:

• Worked closely with data scientists, product managers, and business stakeholders to refine model algorithms and improve data processing efficiency for both HDAM and sub-minute applications.
• Continuously monitored and optimized system performance, implementing updates to maintain high accuracy in demographic predictions, and subminute tuning data in response to changing viewing patterns and data sources.

1. Experian Common Homes and SRLD Processing:

• Developed data pipelines for Experian-matched common homes and Synthetic Respondent Level Data (SRLD).
• Implemented matching algorithms and synthetic data generation techniques to enhance audience measurement accuracy.
• Optimized processing of large-scale datasets, handling millions of records efficiently.

2. PPM (Portable People Meter) Fusion:

• Designed and implemented data fusion algorithms to integrate PPM out-of-home viewing data with in-home panel data.
• Developed matching processes to fuse PPM donor panelists with National TV recipient panelists based on viewing behavior and demographics.

3. Big Data Processing and Analytics:

• Utilized Apache Spark and Hadoop to process and analyze large volumes of viewing data across multiple systems.
• Implemented scalable solutions to accommodate growing data volumes and increasing granularity of audience measurement.