Shikhar Nigam

Performed EDA on Amazon customer data and reviews given by them, build Tableau dashboards to visualise top customers, customers who are likely to churn, top-rated products, least-rated products and the trends in the sales of the products.

Made a classification machine learning model using NLP feature extraction techniques like TF-IDF which segregated reviews into positive, negative or neutral reviews based on the sentiment of the review.

By using unsupervised machine learning… Show more

Performed EDA on Amazon customer data and reviews given by them, build Tableau dashboards to visualise top customers, customers who are likely to churn, top-rated products, least-rated products and the trends in the sales of the products.

Made a classification machine learning model using NLP feature extraction techniques like TF-IDF which segregated reviews into positive, negative or neutral reviews based on the sentiment of the review.

By using unsupervised machine learning clustering techniques like K-means and Hierarchical clustering, identified 3 groups of customers (highly satisfied- around 28% of customers, satisfied - 52% and dissatisfied - 20%) based on the reviews given by them to prevent the churn of dissatisfied customers.

Also identified 3 groups of products( best - 32% of products, good - around 41% and average - 27%) based on the reviews given to them to improve the satisfaction level of products with lower ratings by giving necessary inputs from feedback.

Predicted the future sales of Amazon products to properly maintain the inventory using time-series analysis. Show less

Analyze customer visiting patterns, how much money they’ve spent, and which menu items are their favorite. This will help deliver a better and more personalized experience for loyal customers, whether he should expand the existing customer loyalty program

Predicted Grade of houses using various supervised machine learning classification models(logistic Regression, Gaussian Naive Bayes, Decision Tree Classifier, and Random Forest Classifier using Bagging, Boosting, Stacking) with highest accuracy of 89%, to simplify house search for prospective buyers and help agents quickly narrow down the house according to customer needs.

Applied unsupervised clustering algorithms (Kmeans,DBSCAN,Hierarchical) with highest silhouette score of 0.40 and clustered customers into four categories. Target customers having high income as well as high spending scores, Careful with high income and low spending scores, Careless having less income and high spending scores and Sensible with less income and low spending scores.

Using MySQL estimated the overall profit gained by each movie along with popular actors so that store can stock their inventory to increase profits.

Created a shopping cart program in Python which accepts items from user and display the total bill of shopping along with an updated inventory of the shop.

Used various Supervised Machine Learning Regression Models (Linear Regression, Ridge CV, Decision Tree Regressor, and Random Forest using Bagging, Boosting, and Stacking) with highest r2 scire of 0.55 to predict the sales of the store to boost the sales and expand the business.