Swapnil Rai Rai

Managing a team of data Scientists to help one of the biggest US retail company to price their onsite banners by using various analytical models

key roles & responsibilities

• Tracking the performance of the various onsite banners in ad manager by using CPM ROAS and historical impressions
• Building robust statistical models to predict the right CPM for each banners by using fb prophet based forecasting models
• Responsible for managing a team of more than 5 Data scientists and data engineers to drive day to day project deliverables
• Acting as scrum master to drive the agile based project deliverables in Jira
• Responsible for up skilling mentoring of the team member's through skill mapping right mentorship
• Responsible for driving high client pulse score through effective management of the project

Impact

• Increased revenue for customer by optimizing the pricing of the onsite banners through ensemble based forecasting models
• Conceptualized a web based tool that helps monetization of the banners by providing a centralized booking system that has ability to reserve the banners in a one stop solution

Project 1:

• One of the largest CPG company were interested to figure out the optimal spends allocation and quantifying the impact of marketing channels and understand their MMM Maturity
• This project involved a detailed approach to understand the client requirement and gather the information around their current Marketing effectiveness maturity
• The CPG was interested to developed MMM Model to quantify the impact of the short term, long term, effect of various Marketing channels along with macroeconomic factors
• Responsibilities:
• Developed the outline for Data integration from various ATL and BTL marketing spends
• Developed an EDA plan for univariate and multivariate analysis
• Developed a statistical framework (log-log elasticity model) for feature engineering and variable transformations
• Implemented GLM based Regression models to quantify the impact of various Marketing channels using appropriate Lags and ad stocks for various Marketing channels
• Exploring the Effectiveness of Bayesian regression
• Model validations and interpretation of the result for Business presentation
• Ad hoc analysis based on key ATL Marketing channels using TV ads to understand the impact of various campaign and optimization
• Implementation of the K means clustering techniques to segment the key profitable channels and campaigns based on GRp’s and spends

Project 2:

• Forecasting impressions for world largest home improvement retailer in USA
• Ensemble forecasting algorithm to select the right model for forecasting pageviews for more than 5000 taxonomies,
• Developed and conceptualized the right forecasting algorithms (ETS, auto Arima, SARIMA, UCM) to be used various taxonomies thereby increasing the model accuracy by 20%
• Extensive use of GCP to query data from various sources
• Developed predictive modelling algorithm to identify the key driver of impressions, implemented regression based models to identify the right type of banners driving impressions

• Health economics complex data analytics
• Building Network meta analysis models in R and python for indirectly comparing various studies
• Building complex statistical models such as cox regression models for survival analysis
• Helping team to test various hypothesis by using statistical tests such as paired t-test, chi-square test etc

Project 1:

• One of the largest kidney dialysis headquartered in USA wanted to understand the key drivers of their Net promoter score
• This project involved a detailed approach to understand the data collected by client through surveys and questionnaires and analyze present the key business insights to the client which would ultimately help them in increasing their net promoter scores and better manage their client servicing team

Responsibilities:

• Data integration from various sources to create analytical data set
• Performing the data cleaning, sanity checks and descriptive statistics and correlation
• Developed the statistical framework of for the data analysis and EDA
• Implemented Principal component Reduction to analyze the set of questioners which falls together
• Performed clustering (Means clustering) to segment the profitable customers in different regions
• Implemented NLP for open end questionnaire through text mining, text clustering using topic modelling
• Worked on the translation of the insights to Business presentations for client
• Project 3: Internal product Development based on various forecasting techniques such as Exponential smoothing, ARIMA and UCM
• Build and conceptualize the advance projection-based Methodologies such as ARIMA, Exponential smoothing for the internal web base product development
• Integrating and designing the User Interface design for the projection-based methodologies
• Responsibilities
• Design and integration of the Risk analysis tool based on Monte Carlo simulation for internal product
• Implementation of complex forecasting models such as “ARIMA: Robust ARIMA and exponential smoothing methods in R
• Lead a team of Data scientist to design the robust user interface for the tool and implemented robust Auto Arima techniques to increase the accuracy of the Models
• Health economics outcome Research:
• Implementation of the Advance statistical model for Real world Data analysis
• Bayesian based regression techniques in SAS and R using various ML Models
• Network meta-analysis using R and SAS based tools to generate the statistical indirect evidence out of the Real-world data
• Implementation of the complex regression-based Models for Real world Data a
• Calculation of the sample size for the Retrospective trail
• Implementation of the Advance statistical methodologies such Chi-square test, Paired t- t test, Anova and other non-parametric test in SAS and R

Project 1:

• Implementation of Data driven Trail Execution plan for analyzing real time clinical Research Operations efficiency:
• Designing a excel based Dashboard for tracking the real time clinical operations data obtained through various CTMS
• Implementation of risk-based flagging for various Clinical sites based on the various complex models

Certifications:

• Primary Market Research based study to analyze the role of the Unified "Communication in Pharmaceutical Product Life Cycle Management”