Akshata Patil

[email protected]
571-223-4948

Data Science and statistician professional, creative thinker, and analytical person. Able to distill high
performant solutions from data to drive business strategy. Versatile, purposeful, and meticulous
professional in Data Science and programming. Experience in Machine Learning and Data Mining with
large Structured and Unstructured datasets, performing Data Acquisition, Data Validation, Predictive
modelling, and Data Visualization. Experience in time series foresting, text mining,and deep learning.

Summary
 Skilled in using dplyr, ggplot2, Pandas, NumPy, Matplotlib, Seaborn and sklearn in R and python for
performing exploratory and predictive data analysis
 Experience with Data Analytics, Data Reporting, Ad-hoc Reporting, Graphs, Scales, PivotTables.
 Familiarity with developing, deploying, and maintaining production models with scalability in mind.
 Exposure to working on cloud platforms like MS Azure – Machine Learning and Power BI.
 Proficient in building dashboards in Tableau, MS Power BI, ggplot2 and R shiny.
 Provides business expertise and supports the development of models and analysis to provide the
organization with insights.
 Drives the analytics roadmap proactively by identifying opportunities in the data based on the business
priorities working with all divisions.
 Proficient in development methodologies such as Agile, specifically Scrum methodology.

Technical Skills
Analytics: R, Python, Tableau, Power BI,SPSS Modeler, XLSTAT
Data Science: Generalized Linear Modeling (GLM), LogisticRegression, Decision Trees, Naive Bayes,
Random Forest, Gradient Boosted Models (GBM), SVM, Neural Networks, CHAID, PCA, K-mean
clustering, Discriminant analysis, ARIMA, SARIMA, Holt’s methods, Weighted Moving Average,
Hypothesis Testing,Sentiment Analysis, Topic Modeling and Named Entity Recognition.
Databases: Azure SQL Database, MS Access 2016, SQL Server 2017, PostgreSQL, and MySQL
Cloud Platforms: Microsoft Azure Machine Learning Studio, IBM SPSS Modeler
IDE: Jupyter Notebook, Spyder, Eclipse and Google Colab.
Certifications: Microsoft Azure ML – Data Science Essentials, IBM SPSS Modeler data analysis for
Business partners v2 

Education
JUNE 2012
Master of Computer Applications/M S Ramaiah Institute of Technology, Bengaluru,India
JUNE 2009
Bachelor of Computer Applications/Karnataka University, Vijayapura, India

Professional Experience
SEP 2019 –TILL DATE
Data Scientist – Sales Forecast/ Avanos Medical, Alpharetta, Georgia
Worked as a Data Scientist for a leading medical technology company focused on delivering clinical
medical device solutions. Avanos provides different solutions ranging from Respiratory Health, Pain
Management Solutions, Digestive Health, IV Infusion, and Surgical Pain. Collaborated with a small team
of Data scientists and analysts where we created numerous sales forecasting models from Avanos’s
historical data hosted on Microsoft SQL server to estimate sales forecast for each Business and Group
unit. Forecasts help sales ops with territory and quota planning, supply chain with material purchases
and production capacity, and sales strategy with channel and partner strategies.

 Queried databases with pyodbc in Python and loaded the results into Pandas Data Frames.

 Visualized data and performed exploratory data analysis (EDA) with Matplotlib/Seaborn in
Python.

 Endeavored multiple approaches for predicting month ahead sales demand with Python,
including exponential smoothing, ARIMA, Prophet, TBATS, and LSTM.

 Performed feature engineering with the use of NumPy, Pandas, and FeatureTools to engineer
time-series features.

 Evaluate the performance of predictions using the root mean squared error (RMSE)

 Coordinated with the sales team to understand the problem and ensure our predictions were
beneficial.

 Participated in daily standups working under an Agile Kanban environment.

 By getting down to the device/product and customer level, forecasting accuracy is improved by
better understanding products and customers.

JUN 2018 –AUG 2019

Data Scientist – Customer Churn Analysis/Deutsche Bank, Atlanta, Georgia
Customer Churn, also known as customer attrition, customer turnover, in the loss of clients or
customers. Bankoften use customer churn analysis and customer churn rates as one of their key
business metrics because the cost of retaining existing customers is far less than acquiring a new
one.This project focuses on the behavior of bank customers who are more likely to leave the bank
(i.e. close their bank account).

 Set up architecture for integration between Microsoft Azure Machine Learning Studio and
Dynamics 365 Customer Insight.Inject historic data from Dynamics 365 Customer Insights into
Azure ML Studio using web service API.
2
  The goal of this project is to find out the most striking behaviors of customers through
Exploratory Data Analysis and later use some of the predictive analytics techniques to determine
the customers who are most likely to churn.

 Undertook several classification techniques like Gradient Boosting, Random Forest, Logistic
Regression, AdaBoost and K-Nearest Neighbor (KNN)to predict the probability of a customer is
likely to churn.
 Performed optimization techniques like Cross-validation and Hyper-parameter tuning to
improve the accuracy of the classifiers.

 The end results are shown using Power BI dashboards.

 This churn model helped business to identify their loyal customers to conduct loyalty program.
This unique approach to rewards has helped generate record levels of client satisfaction and
retention with members. Members report more than 85 percent satisfaction with their banking
relationship and have a retention rate of 99 percent.

MAR 2016 –APR 2018

Data Scientist – NOKIA - IRM/ Wipro Technologies, Bengaluru, India
Demand forecasting is a quantitative aspect of human resource planning. It is the process of estimating
the future requirement of demand (man working hours) based on their competency which is most
granular level w.r.t CT (customer team), country, project role, etc. Our task is to build a pipeline that
automatically provides monthly forecasts for the next 6 months based on historical data having a
minimum of 50 months for each competency w.r.t CT (customer team), country, and project role.

 Designed and defined advanced analytics approach for NOKIA IRM.

 Set up architecture for integration between MS Azure ML and Dynamics 365 PSA.

 Inject historic data from Dynamics 365 PSA into Azure ML Studio using web service API key.

 The main assumption in analyzing time series is that the successive values of a variable
represent consecutive measurements of equally spaced time intervals. Endeavored multiple
approaches for predicting monthly skill-wise demand with R, including ARIMA (autoregressive
integrated moving average) model, ETS (Exponential Smoothing), and STL (Seasonal Trend
Decomposition).

 Evaluate all models using MAPE (Mean Absolute Percentage Error). Consider the best
aggregated MAPE out of the models.

 Generate forecast results in the form of output data structure which Dynamics 365 PSA accepts.

 Finally deploy an experiment or model as a web service, which can be consumed in two modes:
RRS (request-response service) and BES (batch execution service)

3
FEB 2015 –FEB 2016
Sr Analyst–Text MiningEmployee Perception Survey/ Wipro Technologies, Bengaluru, India
Employee Perception Survey is conducted annually in Wipro to get feedback from its employees about
the organization. To analyze employee feedback comments available in the form of unstructured text
data. 16 crucial drivers were identified from the employee comments like Rewards, Role, and Customer
Focus, etc. that further provided insights on employee satisfaction.

 Built taxonomy around 16 drivers using IBM SPSS Text Analytics for Surveys from scratch.

 Reporting to the Senior Vice President- HR regarding the latest updates and developments of
the project.

 Conducted exploratory data analysis on text data.

 Preprocessed raw text data by removing URL, removing all irrelevant characters (Numbers and
Punctuation), Convert all characters into lowercase, Remove the words having length <= 2.

 Performed tokenization and Lemmatization using the spaCy library.

 OneVsRest multi-label classification algorithm is built which accepts a binary mask over multiple
labels. Naive Bayes, LinearSVC, and Logistic Regression classifiers are used to predict multiple
labels for instance.

 Built sentiment score model for each driver to identify employee’s positive and negative
feedback.

 Web crawling for online data using Alterian SM2 for Glassdoor and Indeed Employee review
data. And performed competitive analysis on online data of employee reviews.

MAR 2014 – JAN 2015

Sr Analyst– HR Analytics/ Wipro Technologies, Bengaluru, India
Wipro is a big enterprise with 150K employees across the globe in IT / IT enabled services. Attrition, if
happens, affects the company services to the esteemed clientele adversely and hence affecting CSAT.
Organization wants to reverse the trend of attrition by addressing the causes of employee unhappiness.

 A wide range of employee data is collected from central HR team like employee profile data,
leave data, manager data, appraisal data, salary data and training data.

 Queried databases with RPostgres in R and loaded the results into Data Frames.

 Applied dplyr in R to manipulate data and engineer features.

 Built an advanced predictive model using techniques like random forest and logistic regression
in R to explore the complex non-linear relationship between Attrition and its causal drivers. Also
using a scoring algorithm, the list of potential attritors generated.

 The model is evaluated by confusion metrics, you are cross tabulating the actual results with
predicted results. Another way the models are evaluated using AUC. The higher the AUC the
better.

 Finally, generate the report of model accuracy and statistical inference of the model.

4
  Extract and transform data required for Visualization tools.

 The attrition model helped the business minimize the cost of new talent acquisition based on
the employee profiling and company requirements, analysis, and assessment of the loss in
expertise and skillsets, it also helped in preparing contingency plans based on the insight and
foresight provided by the model.

JULY 2013 –APRIL 2014

Project Engineer –Social Listeningand Text Analytics/ Wipro Technologies, Bengaluru, India
Lloyds Banking Group is one of the UK's largest financial services organizations, with 30 million
customers. Was part of advanced analytics team responsible for providing Social Media Analysis report
that contains analysis on text data using different machine learning models for classification and
sentiment analysis based on the different services they provide. Identified insights and provide
recommendations of actions for Present market, Product analysis, Demographic analysis, Platform
analysis, Feature analysis, customer retention analysis, Competitive analysis, Campaign analysis, etc. To
know about the customer interest and attrition reason about the product and service.

 A range of data is collected from various sources like forums, blogs, Twitter, Facebook, open
web, etc. using Alterian SM2.

 Extracted features from the text data using TfidfVectorizer.

 Performed multi-label classification approach for classifying customer reviews from different
social media segments into categories based on their products and services. Linear SVC
performed well in this scenario.

 Generated Word Cloud for most talked-about services and products.

 Customer retention analysis – Identified key factors driving customers who are willing to switch
(Preferred Banks) and key factors driving customers who are willing to stay.

 Did Event Analysis – ATM stopped Working.

 Creation of dashboards in Tableau for the result of models built in Python.

 This helped client continuously refine and improve their strategy to improve your marketing
efforts, stay on top of emerging trends, and offer your clients proof of your results.

AUGUST 2012 –JUN 2013

Project Engineer – Demand Forecasting–Central Workforce Management/ Wipro Technologies,
Bengaluru, India
Wipro is an Indian multinational corporation that provides information technology, consulting, and
business process services with 1.25L employees. Wipro having a large workforce always faces difficulty
in managing the workforce demand. This model helps in optimizing the hiring of resources by skill and
by suit-id based on the respective monthly forecasts Intern helps in streamlining of business process for
the hiring of resources.

5
  The demand forecasting model helps to predict the skill-wise (top-level) demand for resources
based on historical data and to predict the suite-wise (hierarchical) demand for resources based
on the skill-wise forecast.

 Endeavored multiple approaches for predicting monthly skill-wise demand with R, including
exponential smoothing, ARIMA, Holt’s methods, and Croston.

 Incessantly validated models using a train-validate-test split to ensure forecasting was sufficient
to elevate the optimal output of the number of resources (mapped skill-wise) to meet demand.

 Coordinated with the central workforce team to understand the problem and ensure our
predictions were beneficial.

 Forecast capacity vs. demand from multiple perspectives, i.e., by role, department, location,
skills, etc helps you identify short or long-term shortage/excess of resources ahead of time.
Bridge the gap with appropriate resourcing treatments. Take corrective actions, i.e., retrain
current employees and hire recruits early on. Provide shadowing opportunities to bench
resources.

Susan

Sr. Bench Sales Recruiter

Desk: 571-223-5948  Ext (105)

Fax: 571-291-4522

Email: [email protected] 

LinkedIn:  linkedin.com/in/susan-s-838a1719b

44081 Pipeline Plaza Suite 315 Ashburn,VA 20147

www.pvkc.com