Nischal HP

Read this first

Neural Networks - The first cut

As I get ready to do my first solo workshop on Deep learning in the field of NLP, my anxiety levels are extremely high and I am definitely excited.
The anxiety levels have made me read numerous papers , blog posts and definitely some code. I feel I have made a small progress in understanding the subject and this makes me feel a little comfortable.

Deep learning is really not too hard to understand if you think about it from a real world angle. What we are really trying to do is to build systems that can emulate the human mind.

To explain neural networks, I will take a an real world example and then we will see how they are represented in an NN world. Let us take the simple example of catching a ball when it is thrown at us. If you ask a small kid how he does it , you will hear him say well you throw the ball and I catch it. If you try to look a little closer , there are so many

Continue reading →


Pycon India 2015 - Dev Sprint - Data Science/ Visualization

Dev sprint as we had know, is a day that is dedicated to solving bugs of a project or sometimes working on building a new project itself. Never having done a dev sprint before, we chose to bridge the gap between performing data science in python and a way to visualize them in the world of javascript.

So we all know D3 is amazing for visualization and we wanted to do what Vincent was doing for Vega and make it compatible for Vega 2.0. Atleast that was the idea. We did some planning prior to the dev sprint and it really helped us to understand which parts each of us would be mentoring

There were 4 of us [Amit Kapoor] (www.twitter.com/amitkaps) , [Bargava](www.twitter.com/bargava), [Raghotham] (www.twitter.com/raghothams) and [Me] (www.twitter.com/nischalhp) who were conducting this dev sprint and how we achieve our dev sprint was pretty open ended. We started the day at around 10 in

Continue reading →


NodeBoat - Super cool boat building!

Well, I am here in singapore and there is a JSConf happening. I have never been a JS guy but I have come to the understanding that JS is really powerful. Why? I am still figuring out

So I am completely pumped up to attend my first JS conference and suddenly I receive a mailer which said NodeBoat workshop the first day using stream.io and node.

I loved this, mainly cause i am a hardware crazy guy and bringing both the worlds together is just brilliant.

The day started with the awesome organisers giving us spark.io cores. We claimed our cores and based on the pin configuration and charts given to us we built the circuitry. Sounds simple right? Well there are certain things you learn only when you get hands dirty, a simple reverse pin assembly fried our battery pack and then you understand, WOAH this is not as simple as it looks. You really need to pay attention when you are building

Continue reading →


Setting up FISH

Well, I have recently started using a MAC and I am loving the feeling. I almost never use anything that is set by default.

I am currently running on iterm+fish.

Getting iterm and fish working is like a walk in the park. Its easy.

So if you want to change your shell and make fish as default you have to use
chsh -s <path to your new bash>

The moment you have fish running, you will fall in love with it. There is color highlighting, intellisense , autocompletion. Its nice.

But the very next problem you end up having is:

Setting up environment variables

So lets say you need to add a new path to the already exisiting PATH variable in fish, so how do you do it.

its actually very simple but a little hard to find on the internet(maybe i dont know how to search).

The steps to do this are as follows:

1) Create a config file in the following location ~/.config/fish/config.fish

2) Open

Continue reading →


Tweetlyze- Twitter Data Extraction

I started off my MS project on trying to understand Twitter trends and predict the sectors that are affected by the trend , thereby helping people make better business decisions.

I had to throw this idea away as it required a lot of groundwork to be done in understanding various sectors and what can be affected and what not.

But I was sure i wanted to do something with Twitter trends , so i thought why not build a trend summarizer. Trends are so abstract sometimes that it is really hard to know what they are actually talking about. I wanted to solve this problem.

But to solve that problem , I was in need of huge amounts of twitter data related to trends and with respect to time. So I then realized there were two problems I needed to solve:

1. Extraction framework to get Twitter Data      
2. Using this data to build a summarizer.

This post is about all the decisions I took in

Continue reading →