About Me
I work as a decision scientist at Musigma Inc., the world's largest pure-play analytics company. I studied computer science and engineering at Vel Tech Institute of Science and Technology, where I received my bachelors degree.
Analytics Projects
Research Papers
Certifications
Awards
Overview
Summary
Decision Scientist with 1+ years of experience designing and developing analytical and statistical solutions for Fotune 500 clients - Starting from defining the problem to delivering the final consumption framework.
Education
Bachelor of Technology
2015 - 2019
Veltech Dr. RR & SR Technical University, Chennai
Majored in Computer Science and Engineering. GPA 9.26/10.00
Data Structures and Algorithms Specialization
Senior Secondary Certification
2011 - 2014
City Montessori School, Lucknow, India
Indian Certificate of Secondary Education(ICSE)
XII - 86% | X - 92%
Professional Experience
Decision Scientist
July 2019 - Present
Mu Sigma Inc., Bangalore, India
- Understanding client’s business requirements to define problems; create end-to-end data driven solutions, generate insights and present it to the Business & Analytics teams of customers
- Implemented advanced predictive models(Regression & Classification) using Python/Py Spark for analysis, SQL for data management; PowerBI /Tableau for dash boarding
- Designing optimization techniques and building solutions for $12M+ portfolio
- Deploy and track machine learning models in a production enviroment
Projects
As a decision scientist, I have worked on some of the highly statsitical and impactful modeling projects. A few of them are briefly outlined below.
Merchandising Planning and Optimization
Russian Petcare Manufacturer
- Increased merchandizing efficiency by creating optimized merch visit plans to maximize sales uplift and ROI
- Clustered stores based on various factors and estimated the impact of visits after isolating the effect of seasonality and promotions
- Identified the most optimal plan for 40k outlets given a certain budget and maximizing either NSV or ROI based on business requirement
- Built a reporting and optimization tool using PowerApps and PowerBI which allows user to build optimized merchandising plans, report results and execute the plan on the field.
- Impact of ~3% incremental sales uplift(+$6.2MM) and sales force expenditure reduced by 24%($0.5MM) – and savings of 960 hours/quarter effort
Marketing Mix Model and Prioritization Engine
Energy Bar Manufacturer
- Created a marketing mix/driver model for attribution of sales uplift to different marketing levers like price promotions, social media, TV advertisements, In-Store sampling, features and displays with MAPE less than 20% using ElasticNet regression
- Designed an optimizer to create the best marketing plan which takes into account the investement associated with each promotion channel and provides recommended investement across each channel to maximize ROI over the limited budget available for promotions
- Built a prioritization engine which recommends the stores and regions sales reps need to focus on for maximum uplifts, reducing $390k in in-store sampling investments and increasing sales uplift by 7%
Commercial MARIO Reporting
Russian CPG Manufacturer
- Built and automated the creation of business data mart – which was the single source of data to upsurge reporting of SAP Atlas based reporting metrics for all manufacturing plants across Russia
- Implemented a data quality management layer to the harmonized data source to improve the quality and scope of data driven decisions; that was subsequently leveraged to refresh 20+ dashboards relying on SAP Atlas data - resulting in 80% man-hours savings
- Designed and developed user friendly flexible and interactive PowerBI Dashboards to track and assess the performance of supplier contracts across Vendors/Materials/Plants
Machine Learning Research
I have been fortunate enough to publish multiple research papers. Some of them are highlighted below.
Optimizing Stock market prediction using LSTM networks
This paper proposes using a LSTM and auto-encoder based deep learning model to predict stock market price. The idea is to forecast stock prices and allocate stocks to maximize the profit within the risk factor range of the stock buyers and sellers. A YouTube video with a detailed explanation and presentation can be found here.
Social Media Data Analysis: Twitter Sentimental Analysis
This paper talks about sentiment analysis of tweets on the Barcelona terror attacks and observe, examine and analyze how people criticize a situation either by expressing their aggression against terrorist or supporting the victims. The tool used is R - a statistical modeling program.
Feature Selection Effects on Classification Algorithms
This paper compares the effect of Feature selection on various Classification algorithms. The vitality and vulnerability of K Nearest Neighbor, Naive Bayes and Support Vector Machines is examined. Guidelines to perform feature selection on categorical dataset is also explained laconically.
Ganesh Kalyanasundaram
Strategic Consultant
Sonal Singh I can recommend as a person with great proficiency and deep experience of solutions. Sonal is a self motivated and wise person but also an inspiring perfectionist. Creative strategist with great interpersonal skills. She is an honest team player. If you ever need someone to deliver under pressure, no slip-ups, just results, Sonal is your go-to person!
Syed Misbah
Data Scientist
Sonal is a rare combination of brains and diligence. Smart, dedicated and has a highly analytical bent of mind. Approaches every problem with crystal clear clarity. Sonal has been a pivotal member of some of the biggest analytical projects in Mu Sigma. Her background as a machine learning researcher gives her an added advantage.
Shubham Lahoti
Decision Scientist
Sonal is a very dedicated and hard working person, who aims at perfection in whatever she does. She has a great ability to grasp and understand critical concepts quickly and apply them in solving business problems. Apart from her problem solving skills, she has got a great oratory prowess, which helps her put her thoughts across very clearly. Having worked with her in a project for a brief period of time, I have closely seen her multitasking, yet excelling at whatever she does, which is a rare quality. I would highly recommend Sonal
And a little bit about me!
Here are a few things that you might be interested to know about me.
MBTI
My Myers-Briggs Type Indicator(MBTI) is ISTJ [Introverted Sensing Thinking Judging]. The favorite people I supposedly share my personality type with are Queen Elizabeth, Warren Buffet and Satyajit Ray.
Quora
I write on Quora whenever I feel like I have something new to add, usually about India, life, code and education. Although I am not usually active these days because of work pressure.
Kathak
I am a trained Kathak dancer and have been on it since the age of 4. Kathak originated within temples as a storytelling device for portraying the epic tales contained within the scriptures. Poetry was combined with rhythmic movement to aid in the worshipful storytelling.
Foodie
I am an avid foodie - and love to cook food, rather than just eating it :P
I can cook pan-asian, Indian and continental cuisines. My favorite dish has to be a tie between hot steamed chicken momos and biryani.
Contact Me
Want to reach out to me? Contact me using the form below
Social Profiles
Email Me
sonalsingh2606@gmail.com
Call Me
+91##########