Vallabha Mohta
Contemplative analyst, inspired by logic and meaning
Work Experience :
Forecasting Stock Prices | Sight Spectrum
May - August 2019
-
Using Nifty50 stock price data of 5 years at a granular level of 1 min and 1 hour to forecast price movements for different time intervals of 15 mins, 30 mins, 1 hour, etc.
-
Preprocessing the data and building predictive statistical time-series models and machine learning models using R, MS Azure, BigML.
-
Conducted original research to create a model to explain perceived vulnerability to loss using a secondary dataset from a previous psychological study.
-
Performed statistical analyses (correlations, t-tests, ANOVA) using Stata to build the best-fit regression model using the given variables which explained 41% variation in the dependent variable.
-
Wrote a research paper compiling the findings.
November - December 2018
Manufacturing Quality Analysis & Simulation
-
Examined product and quality records using MS Excel and Stata to find correlations, conduct ANOVA and Chi-Square tests, and build regression models to analyse demands and orders, and to determine the factors that affecting quality scores and complaints.
-
Built and validated a simulation model of the current production system of the manufacturing unit using Simul8 and provided expansion suggestions.
November 2018
Multiple Regression Analysis
Customer Savings & Investments,
Operational Analysis & Insights Team | M&G plc. (Prudential UK)
October 2019 - Present
-
Collaborated with senior management teams to identify and interpret patterns and trends in data for prioritising business requirements and locating opportunities for process improvements.
-
Curated and analysed big data across SQL/Oracle/SharePoint, conducted
descriptive and prescriptive analysis, and built performance dashboards to facilitate audit, reporting, and oversight.
-
Designed the data and reporting infrastructure from ground-up for multiple
datasets across customer interactions, workflow, service experience, and
expectations to provide actionable insights into business KPIs.
• Built and automated extraction and transformation processes for Marketing, Research, and Digital Adoption campaigns; achieved 100% of target registrations and reduced processing time by 80% each week. Produced regular MI to evaluate the success of campaigns and optimise ROI.
• Devised data extraction pipelines and work-stack forecasting models on Databricks and Python for predictive analysis; presented results on interactive Power BI dashboard with RLS to help in effective capacity
planning.
• Developed and revamped existing processes from SPSS to SSIS; improved
run-time by 70% and reduced manual intervention by 40%.
Customer Journey & Conversions Analysis | VisitScotland
February - April 2019
-
Analysed web data (tourism industry) using Google Analytics to understand customer journeys and interactions.
-
Provided insights on customer analytics, highest traffic sources, market segmentation, and exit pages that led to conversions.
-
Communicated actionable recommendations to shareholders using visualisations from Tableau and MS Excel.
-
Cleaned the stores' data and build supervised regression (Random Forest, MLP, Linear Regression) and classification (MLP, J48, Logistic Regression) machine learning models (with Ensemble method - Bagging) to predict profits and classify the performance of the stores owned by a retail company to assist in determining profitable stores and preferable locations.
-
The best-fit models after tuning of the hyperparameters were MLP-with Bagging (explaining ~78% variation in the dependent variable) and Logistic Regression (~77% of prediction of all classes were estimated correctly by the model).
-
Used Weka software and R for this analysis and compiled a thorough report explaining all the steps.
Predictive Retail Analytics
January 2019
December 2018
Data Exploration using Python
-
Performed data exploration of a secondary dataset containing explanatory variables of every aspect of residential houses, using key visualisations and preprocessing techniques in Python.
-
Used various Python libraries like Numpy, Matplotlib, Pandas, to generate histograms, scatter plots, boxplots, heatmap correlation matrices to visualise the data.
-
Checked for outliers, missing values, and standardised the data to make it suitable for modelling.
