Portfolio

Power BI Dashboard for Retail Business
‍

Introduction
‍
This summary provides a detailed overview of a projectfocused on developing a Power BI dashboard for a retail business operating 12stores across Ahmedabad. The aim of the project was to provide the retailbusiness with actionable insights into various aspects of their operations,including sales, inventory, customer behavior, and store performance. Byleveraging Power BI's data visualization and analytics capabilities, thedashboard aimed to empower the business to make data-driven decisions andimprove overall efficiency and profitability. 

Project Objectives
‍
Centralized Data Management: The project aimed toconsolidate data from all 12 stores, including sales transactions, inventorylevels, and customer information, into a centralized database for analysis andreporting purposes.
Visualization of Sales Performance: The dashboard wouldprovide real-time visibility into sales performance across all stores, enablingthe retail business to identify trends, track sales targets, and make informeddecisions to optimize sales strategies.
Inventory Management and Optimization: The Power BIdashboard would display key inventory metrics, such as stock levels, turnoverrates, and product performance, helping the retail business to streamlineinventory management processes and reduce holding costs.
Customer Insights: By integrating customer data from varioussources, including loyalty programs and CRM systems, the dashboard would offervaluable insights into customer behavior, preferences, and purchase patterns,enabling the business to personalize marketing efforts and enhance customersatisfaction.
Store Performance Analysis: The dashboard would provide acomprehensive view of each store's performance, including sales per squarefoot, footfall analysis, and comparison of key metrics between differentlocations, allowing the retail business to identify top-performing stores andareas for improvement.
‍
Project Implementation
Data Integration: Data from each store's point-of-sale (POS)system, inventory management system, and customer relationship management (CRM) platform were extracted, transformed, and loaded into a centralized datawarehouse.
Data Modeling: The data warehouse was designed, andrelationships between various datasets were established to ensure accurate andmeaningful analysis.
Dashboard Design: The Power BI dashboard was created with anintuitive and user-friendly interface, featuring visually appealing charts,graphs, and tables to present the data in a concise and informative manner.
Report Generation: Customized reports were developed toprovide deeper insights into specific areas of interest, such as sales byproduct category, customer segmentation analysis, and store performancecomparisons.
Automation and Data Refresh: Automation processes wereimplemented to ensure the dashboard and reports were regularly updated with thelatest data, allowing stakeholders to access real-time information.

Project Outcomes and Benefits
Improved Decision-Making: The Power BI dashboard providedthe retail business with valuable insights and actionable information to makeinformed decisions regarding sales strategies, inventory management, andcustomer engagement.
Enhanced Operational Efficiency: By streamlining dataanalysis and reporting processes, the dashboard eliminated the need for manualdata manipulation and significantly reduced the time spent on generatingreports, allowing the business to focus on value-added activities.
Increased Sales and Profitability: The ability to monitorsales performance, identify top-selling products, and analyze customer behaviorempowered the retail business to optimize sales strategies, personalizemarketing efforts, and enhance customer satisfaction, ultimately leading toincreased sales and profitability.
Store Performance Optimization: The dashboard's storeperformance analysis capabilities enabled the retail business to identifyunderperforming stores and take corrective actions, while also recognizingtop-performing locations and replicating successful strategies across otherstores.
Scalability and Flexibility: The Power BI solution provideda scalable platform that could accommodate future business growth, allowing theretail business to easily incorporate additional stores or expand the scope ofdata analysis as needed.
‍

Bangalore's Swiggy Sales Data Analysis
‍

Swiggy is an Indian online food ordering and delivery platform. Founded in July 2014, Swiggy is based in Bangalore and operates in 500 Indian cities as of September 2021. This project was a part of my internship at iNeuron Pvt Ltd, Bangalore. In this project I was supposed to analyze Swiggy's daily transactional data provided by the company and create business report along with the dashboard which represents the findings of analysis.
In order to perform this analytics project I have performed following steps:
- As the data was directly fetched from Swiggy's datalake, it was in a raw form and in order to transform it into usable form I performed data cleansing process and for that I have used Excel, Python - Programming language and Mito which is spreadsheet tool.
- After performing the transformation process on this dataset I have generated relationships between different parameters and plotted different visuals for creating the business report.
- Along with that I have used Tableau for building the interactive dashboard which involves all the findings from the analytics process.
Interactive Dashboard: Swiggy Sales Dashboard
- Along with this I have also created High Level Document, Low Level Document, Architecture and Wireframe document for better understanding of this project.

Air BNB Business Data Analysis for US and Canada
‍

“Air Bed and Breakfast,” is a service that lets property owners rent out their spaces to travelers looking for a place to stay. Travelers can rent a space for multiple people to share, a shared space with private rooms, or the entire property for themselves. In 2017, the number of adults using Airbnb in the United States amounted to 33.9 million, up from 29 million the previous year. This figure is forecast to reach 45.6 million by 2023. And in Canada based on the 12-month period (April 2016 to March 2017) there were approximately 70,000 Airbnb hosts with over 100,500 listings in Canada generating one-half billion in revenues. Which indicates that this industry is experiencing exponential growth. So, in order to analyze and understand this business thoroughly I was given the opportunity at iNeuron Pvt Ltd to find out insights from the dataset of Airbnb for US and Canada. In which I have performed following tasks:
- Performed Data Cleaning with Excel and Python
- Created relationships by understanding the features
- Plotted relationships for better understanding of data
- Built Interactive Dashboard using Tableau for answering business questions
Interactive Dashboard: Air BNB data analysis dashboard
-  Along with this I have also created High Level Document, Low Level Document, Architecture and Wireframe document for better understanding of this project.
‍

Analyzing the customers’ sentiment on Apple’s new MacBook Air
‍

This research is intended to analyze the public sentiment about the new ultramodern apple’s MacBook Air M2. Apple, a gigantic tech company, which has already excelled in the field of technology by making the computers with cutting edge of technology. Apple has introduced the first ever MacBook in the year of 2008 which was promoted as the world’s thinnest laptop and afterwards apple’s R&D team has revolutionized the word with the next generation laptop series. Now, in the year of 2020, apple has announced their own processor named M1 chip which was 14 times faster than the previous one and users were very satisfied with the performance of this laptop. But this year, apple has announced their advanced version of M1 chip which is called M2. And creators have stated that this chip would be 35% more faster than the previous one. Although, the M1 chip was capable enough to manage the high-performance calculations but customers were curious about this latest M2 and as this MacBook Air M2 is now available for customers so, I wanted to know their sentiment about this product. So, by performing this analysis I would like to summarize the public sentiment about this product and the research question would be: What are the public opinions about the latest MacBook Air M2?
‍
Data gathering and cleaning Process:
In order to collect the tweets from the Internet I have used the Tweepy which is python library for accessing the twitter API and with the help of that we could scrape the twitter. Now, Tweepy has some limitations, like we could not scrape more than3200 tweets and could not scrape tweets older than 7 days. But as we are collecting the most recent tweets for this analysis it won’t be any issue. In this process I have used keyword like: ‘MacBook’, ‘Apple’, ‘M2’, ‘Air’. And along with the tweet text I have scrapped the username, tweet time and hashtag for more information .After the data collection part, I have performed some data cleaning techniques for making the data more insightful and removing the unwanted words (stop words) which could create the hurdle in analysis phase.

Analysis report: Analyzing the customers’ sentiment on Apple’s new MacBook Air
Presentation Deck: Presentation‍
‍
‍

Predicting Football Player's Commercial Value
‍

Prediction of future market players' value is important because it could seriously burden professional football clubs. On the other hand, it allows clubs to gain profit by selling a well-performing player at a high price. But this process can arise one question: Is commercial value truly only dependent on pure playing field attributes? Or is it something else? So, in order to find out this I have analyzed huge football dataset and implemented machine learning model using XGBoost regression algorithm which is ensemble technique for machine learning solution. The data set comprises over 18k entries for football players, ranked value-wise, from most valuable to less. Although the data set relates to the FIFA ’19 video game, its player commercial valuations and the player’s play skills ratings are very accurate, so we can assume we are working with real life player data.
‍
This project involves several data analysis tasks such as data importing, cleaning and normalizing which are shown in the following python notebook document.
Along with that, in order to build accurate machine learning model with most relevant features I have performed feature selection process for given dataset. And then I have utilized XGBoost regression algorithm for training machine learning model.
With this model I have achieved accuracy of 90.5% which is considered quite good for this sector.
‍
Python Notebook: Regression Model For Predicting Football Player's Commercial Value