Overview
Human being are doers! With our creativity and curiosity we find ways to utilize whatever resources we have to build something new, useful and fascinating. Right now what we have at our hands, is a lot, and I mean a lot of data. It is growing at quintillion bytes everyday. How much is that? Well, Ben Walker explains this very nicely here. It is as much as
"10 million blue ray discs, the height of which stacked, would measure the height of 4 Eiffel Towers stacked on top on another."
So naturally we have got to utilize it. This is very exciting because we have an opportunity like never before, to understand our social, economic, political and cultural behaviour, learn from it and make predictions about our future. It's like learning from our history, except that our history is being transcribed in memory on the server and cloud, and it's for the taking anytime, anywhere, by anyone. Data science is infiltrating everything, from businesses making data driven decisions to the mechanics behind social activism, and how movements spread and gain momentum, see here. In order to tackle this gigantic task of extracting information and separating the signal from the noise, a new breed of architects is growing. They have several names and roles, data wranglers, data miners, data architects, data scientist, data analysts, data engineers, and are contributing equally.
Yes, I am super excited to be a part of this community and the reasons above are good enough. But personally there are several other aspects of data science which excite me even more. To begin with, it's a lot of fun and I get to do really cool stuff like this.
I compressed the image by using less number of colours using the means clustering algorithm to choose only 16 colours. So image goes from 24 bit encoding to only 4 bit which is a compression by a factor of 6, and yet we can still make out that it's a picture of me.
Here I analysed the pokemongo tweets, and this is a word cloud of the most popular words which occurred in the tweets. The bigger the size, the more frequently the word appeared in the tweets.
Another thing I love about data science is that it's part science and part art. There is a big technical aspect where one needs to very careful preprocess data according to the needs of the model you are trying to fit, make sure the assumptions in the algorithm are met, and that things are mathematically sound. At the same time there is a lot of room for creativity. Taking the raw data with numbers and text, and transform them into beautiful and meaningful visual graphs to tell their story. There is no set hardwired way to fitting model, it's more about trying different things, and tweaking the hyperparameters to continually improve your model. I want to close with one my favourite quotes which sums up the last sentiment very aptly.
Human being are doers! With our creativity and curiosity we find ways to utilize whatever resources we have to build something new, useful and fascinating. Right now what we have at our hands, is a lot, and I mean a lot of data. It is growing at quintillion bytes everyday. How much is that? Well, Ben Walker explains this very nicely here. It is as much as
‘Essentially, all models are wrong, but some are useful’ George E. P. Box