Python
Statistical & Probabilistic Analysis
This project consists of three case studies, which was analysed using python:
Case 1- using 'Wholesale Customer Data (Store Sales)' data and EDA, analysed the pattern of sales and suggested its marketing implications
Case 2- using 'University Survey Data', analysed the career progression of graduate students
Case 3- using 'Manufacturing Shingles Data', tested the whether the shingles have less mean moisture content (< 0.35 pounds per 100 sq ft.)
Tool used
Excel and Python
ANOVA and PCA
This project consists of two case studies, which was analysed using python:
Case 1- using ANOVA analysed the relationship between 'Salary' and 'Educational Qualification' and 'Occupation'.
Case 2- using the dataset on 'Education - Post 12th Standard', conducted PCA to optimize the variable analysis which further helped in pinpointing the factors affecting the admission of students to college.
Skills and Tool used
Python, ANOVA, PCA, EDA
Bank Clustering Analysis
In this project, I developed a customer segmentation from a given bank data and analysed it to develop credit card promotional offers to its customers.
Case Steps
Following steps are taken to solve this case
1.1 Conducted exploratory data analysis (Univariate, Bi-variate, and multivariate analysis).
1.2 Scaled the variables
1.3 Applied hierarchical clustering to scaled data and then identify the number of optimum clusters using Dendrogram
1.4 Applied K-Means clustering on scaled data and determine optimum clusters. Applied elbow curve and silhouette score to find the optimum cluster number
1.5 Described cluster profiles for the clusters defined.
Skills and Tool used
Python,Clustering, CART, Random Forest, Artificial Neural Networks
Market Basket Analysis
The project involves conducting a thorough analysis of Point of Sale (POS) Data for providing recommendations through which a grocery store can increase its revenue by coming up with attractive combo & discount offers for customers.
Used tableau to conduct exploratory data analysis and visualize the sales pattern
Used KNIME workflow to conduct market basket analysis to gain granular understanding to customer purchase pattern and recommend purchase combos and discount offers on specific products
Skills and Tools
Market Basket Analysis, Exploratory Data Analysis, KNIME, Python
Health Insurance Cost - Consumer Analysis Cost Prediction
The objective of this project is to profile consumers by understanding their health parameters through various data analysis techniques and build a model to estimate optimum insurance cost. The business opportunity lies in accurate estimations of insurance cost while social opportunity in providing suitable insurance cover.
Through exploratory data analysis, 2 groups of insurance payers (high and low payers) were identified and profiled according to their different health parameters. The EDA further helped in identifying the characteristics of these two groups and notable finding was that more the weight of the customer more is the insurance cost.
Different models were built to predict the insurance cost and Gradient Descent technique gave the best predictive result while linear regression helped in identifying important features affecting insurance cost.
Skills and Tools
Cluster Analysis, Decision Trees, Python, Random Forest, TABLEAU
Utilizing the ebook version of the renowned novel 'Pride and Prejudice' obtained from Project Gutenberg as the training text corpus to build a model that can predict next word on the given user prompt. Used LSTM and BiLSTM methods to build the model
Skills and Tools
Pyhton, Tensorflow, BiLSTM, one-hot encoding