Computer Science Graduate with specialization in Data Analytics who loves to swim and ride bike but above all is passionate about "TEXT" and spends most of the time in playing with varities of text data.
During my Graduation at Illinois Institution of Technology, Chicago I have learnt various Data Analytics methodologies and skills like Data Collection, Mining, Preprocessing, Modelling and Visualization. I also have interest in Machine and Deep Learning along with Amazon Web Services.
Real-Estate Analysis for determining the rate of the properties according to the indicators such as vacancy level, rental rate and level of income.
The job profile included following tasks:
- Data Collection & Optimization
- Generating Data Files for an Individual
- Automating Data Feeds
- Customized Reports & Visualization.
Building and Training Neural Network to predict the number of bikeshare users on a given day.
Bike sharing systems are new generation of traditional bike rentals where whole process from membership, rental and return back has become automatic. Data being generated from this is of great importance for reseach and this virtual sensor is useful for keeping a check on the mobility of city by which important events can be monitored in the city.
In this we build a Neural Network by implementing Gradient Descent and Backpropagation and train that network on Bikeshare data for the prediction of number of users on a particular day. The dataset has various attributes such as date, season, year, month, hour, holiday, day of the week, windspeed etc. that drives the prediction.
Using the facebook's "like" data of Bill Gates we found communities by fetching the users he liked and one more hop from that and further using the same logic we predicted links which can be used to suggest new links or to recommend friends.
Realtime tweets were fetched using Twitter Streaming API and thoughts of people about Donald Trump were recorded. We found out about the sentiments in the thoughts to predict the possible communities of people who are in favor or against voting him.
Using the genre of the movie as the parameter we find out similarity between movies. By accessing the ratings user has given to movies we predict and suggest a similar movie to the user based on the tf-idf (term frequency–inverse document frequency).
Movie reviews were fetched from IMDB and sentiment analysis was done using token feature and vectorization. Using our model new reviews were taged as positive or negative reviews and movies were rated depending upon that.
A simulator to find out shortest path between two routers and give out the best path from source to destination by building a connection table. The path structure of the routers were provided and Dijkstra Algorithm was implemented.
Developed a Machine Learning application using Python to identify and analyze Tweets of movie reviews using Twitter API. Sentiments of tweets were detected and were classified into different classes as positive, negative or neutral.
Visualization of the political networks of recent US election candidates. Finding out the common links between the candidates and the reason behind the common link.
