Create your Sourcerer profile
Start
1

Ajay Shewale

ajayshewale

Co-founder@blubyn | Entrepreneur | Machine Learning and Data Science
Mumbai
Commits
27
Repos
5
Lines of code
29045
Following
0
2
Overview
5 repos
Last updated: 2018/12/31 — 17:12:00
3
Languages
5 repos
Last updated: 2018/12/31 — 17:12:00
Python
 
Commits:
18
LOC:
847
HTML
 
Commits:
12
LOC:
27766
Arduino
 
Commits:
1
LOC:
41
4
Technologies
5 repos
Last updated: 2018/12/31 — 17:12:00
Computational Science
3 commits
Deep Learning
3 commits
Machine Learning
3 commits
Data Science
3 commits
Python Web
2 commits
5
Fun facts
5 repos
Last updated: 2018/12/31 — 17:12:00

I'm most productive on

Tuesdays
16% of users
Tu

I'm most productive during

night-time
6% of users
n

I prefer

snake_case

for naming variables

I prefer

spaces

for indentation

I prefer

list comprehensions
6
Repositories
5 repos
Last updated: 2018/12/31 — 17:12:00
#
Repository
Commits
Team
Language
Timeline
2
Sentiment-Analysis-of-Text-Data-Tweets-
7
1
HTML
This project addresses the problem of sentiment analysis on Twitter. The goal of this project was to predict sentiment for the given Twitter post using Python. Sentiment analysis can predict many different emotions attached to the text, but in this report, only 3 major were considered: positive, negative and neutral. The training dataset was small (just over 5900 examples) and the data within it was highly skewed, which greatly impacted on the difficulty of building a good classifier. After creating a lot of custom features, utilizing bag-of-words representations and applying the Extreme Gradient Boosting algorithm, the classification accuracy at the level of 58% was achieved. Analysing the public sentiment as firms trying to find out the response of their products in the market, predicting political elections and predicting socioeconomic phenomena like the stock exchange.
3
DLND-Image-Classification
3
1
HTML
In this project, you'll classify images from the CIFAR-10 dataset. The dataset consists of airplanes, dogs, cats, and other objects. You'll preprocess the images, then train a convolutional neural network on all the samples. The images need to be normalized and the labels need to be one-hot encoded. You'll get to apply what you learned and build a convolutional, max pooling, dropout, and fully connected layers. At the end, you'll get to see your neural network's predictions on the sample images.
4
Django-Deployment
2
1
Python
This is Django Deployment exercise
5
Bad-posture-recognition-and-heart-sensor-monitor-using-Arduino
2
1
Arduino
Embedded system becomes very popular these recent years, especially in health monitor system. Therefore we are working on the application for Bad posture and heart pulse monitoring technology using Arduino, Accelerometer, Ultrasonic sensor, heart sensor.