Thursday 29 March 2018

A Review On Tensorflow Course By LinkedIn


Hey aspirant, let me give a quick overview answering 4 major questions related to course, who can enrol, what language you require and who can benefit from the course?

The course is offered since August 2017, if you remember correctly Tensorflow stable version was be launched after 2017 and the course only concentrate on Tensorflow. As you know, tensorflow is one of the finest ML skill after, Spark ML, Keras, H20 etc.

What is all about the course?

The course is divided into 6 parts. Initial chapters i.e 1 and 2 discuss Tensorflow, installing it, hardware requirements, its competent Libraries, how the data need to be fed to tensorflow and others.

In the 3 chapters, you will find how many types of data importing techniques you can use and which need to be used when. The 4th chapter discusses training a model using tensorflow. The 5th chapter discusses the data visualisation and output using tensorboard.

In the final chapter, it speaks about using tensorflow for development on google cloud.


Who can benefit from this course?

As most of the data science aspirants are a self-learnt data scientist and it becomes essential to know when to enrol this course.

 If you are well aware of Neural Network main terminologies like Nodes, Layers, Epoch, Loss functions, Weights, training and testing models then it is good to enroll the course.

If you are able to write a neural net on your own, understand the above technologies and have a overview how to use it. Then you will find easy to understand this course. Even a intermediate understanding on Neural Net is good for the course.



The course is good for the people who have good knowledge on Machine Learning and Deep Learning. The course does not emphasise on ML/DL and needs knowledge of Neural Networks.

Tensorflow on Python or R?
The course is completely thought on Python and if you use R alone the course will give a complete tutorial on installing python to using tensorflow on python A-Z.

Can I use Tensorflow on R? 

Obviously, you do have a package on R and works similar. Only the data importing technique is different compared to Python

Is it value for money?

LinkedIn is always a economical study technique and the course is filled with right amount of theory and practical but speak intuitive level of Deep Learning. You can also know  

Wednesday 14 March 2018

Indian Telecom Voice Call User Experience & Rating Between June-2017 to Feb-2018


Indian Telecom Voice Call User Experience & Rating Between June-2017 to Feb-2018

Introduction

The dataset is from data.gov.in and can find the dataset here: https://data.gov.in/catalog/voice-call-quality-customer-experience

The data consist of 9 month user rating of 9 popular telecom companies in India. The data collection is between June 2017 to February 2018.

You can see the sample of the data below

##   Month Operator   Rating       Type Year
## 1     6      Jio 3.175126    OutDoor 2017
## 2     6      Jio 3.555710     Indoor 2017
## 3     6      Jio 3.307873 Travelling 2017
## 4     6     Idea 3.952443     Indoor 2017
## 5     6     Idea 3.515842    OutDoor 2017
## 6     6     Idea 4.474982 Travelling 2017

Average rating of Telecom Brands

Now lets find the average rating of all networks

As per the data Jio is leading in rating followed by BSNL and Reliance with the lowest rating

Time Series of Rating

Now lets look at the overall rating of all the companies month wise

We can see the fluctuation in the rating on month on month basis it might be due to customer churn, bad network, customer moving to other location etc.We can see a sudden drop in the rating after December 2017

Overall rating of all brands

The data is been collected in 3 phase. Indoor,Outdoor and Travelling. So lets look at the rating of all the telecom companies.

Conclusion

Jio turns to be a market leader followed by BSNL. We do not see any significant difference between networks but reliance is slightly on a lower side.