Swetha Nandyala

Consumer Creditworthiness Analysis:

Project Description:

It is a machine learning project in the banking domain. It is a banking data which contains information about the status of loan approval. Since bank receives many loans applications on the daily basis. So, the higher management notices that this takes significant amount of time of employees. Since the number of applications for loan approval are also very high that is why they are not able to filter the appropriate candidature for loan approval.

Roles and Responsibilities:

• Developed a Machine Learning model using classification algorithms to aid bank employees in identifying qualified loan applicants.
• Performed Hypothesis Testing, EDA which includes univariate, bivariate analysis, Multivariate Analysis, Feature selection, Encoding, Feature scaling.
• Various classification algorithms like logistic regression, Decision Tree, Random Forest, Ada boost and Gradient Boosting were analyzed to check the predictions.
• Analyzed the performance of models using metrics like Accuracy, Precision, Recall, F1score, ROC-AUC, Precision-Recall Curve.
• Achieved an Accuracy rate of 80% and ROC_AUC score of 75%.

Project Description: Social Buzz is a fast growing tech unicorn that need to adapt quickly to their global scaling process efficiently. Accenture will run 3 months POC and the Data Team will focus on analysis of their content categories that highlights the top 5 categories with the largest aggregate popularity.

Roles and Responsibilities:

• Understand the client and business problem at hand, identify the requirements that need to be delivered for this project.
• Analyze the content categories that highlight the top 5 categories with the largest aggregate popularity
• Understanding the relevant datasets to answer the business problem and performed Data cleaning and Data Modelling
• Built interactive dashboards to bring data into life.
• Prepare powerpoint presentation to communicate findings to the client.

Project Description:

This Projects aims at creating a dashboard for Walmart Retail Data 2015. Generate insights on which age group customers were profitable.

Generated Insights on:

• Built interactive Dashboard that displays:
• Sum of sales by state on a filled map.
• Donut chart for Avg sales for each product category.
• Discrete Line chart for avg sales for each quarter per year and sales per year.
• Dynamic Bar chart Customer age Vs Avg Profit using customer Age as a parameter.

Project Description:

Analyze the US Accidents data 2021 and report which states have greater number of accidents. Analyze the number of accidents according to the timezone, Months, Weekdays, hours. What are the environmental and other factors contributing to the accidents.

Generated Insights on:

• Sates which have highest number of accidents recorded.
• Weekdays, Months, Hours during which the accidents are high.
• Analyzed the effects of environment (sunrise and sunset, temperature)and other factors like traffic signals, Point of Interest on the number Accidents.

Census 2011:

Project Description:

Create a dashboard using the India census 2011. What are the states with highest population and population density. Analyze the population by Gender in each state. Provide details on the number of villages inhabited and uninhabited in each state.

Generated Insights on:

• Percentage of population per state
• Gender diversity in each state.
• The number of villages inhabited and uninhabited in Rural and Urban areas and area in sq km for each state.
• Analyzed the states with highest population density.

Project Description:

Build an application to scrap the data which contains information about the courses offered by Ineuron.

Roles and Responsibilities

• Developed application in python that scrapes the details of each and every course present in the ineuron website.
• It fetches the URLs of the available courses and saves them to excel.
• Then scraps all the details from each of those links and saves the details into database (Mongo DB)
• It fetches the data from database and displays the data in JSON format.
• The application was tested using postman.

Client BMW:

Project Auto Calibration Tool:

Auto Calibration Tool is developed to reduce the manual effort spent in the manual calibration process. Vehicle crash data is collected from various sensors (Acceleration, Pressure, Combo sensors) during a crash event. The data preparation has to be done and the tool has to discriminate between the Must deploy crash and the No deploy crash by combining various features of the data.

Roles and Responsibilities:

• Worked closely with Subject Master Experts to identify and define business requirements.
• Responsible for data collection, validating the ACT tool and giving insights to the counterparts.
• Responsible for providing reporting, analysis and insightful recommendations to developers on key performed metrics.
• Tested clustering models Hierarchical clustering.
• Performed statistical analysis and applied machine learning techniques to the data for identifying potential areas of enhancements in tool.

Result: Achieved manual effort reduction of 30% percent with the tool. Increased productivity.