Hi, my name is Shéhaaz Saif from freezing Montreal, Canada. I am a fourth year Computer Science Major and Psychology Minor from Mcgill University. I like long walks on the beach and Java.
At the moment, along with attending school, I am interning at PTC, it is a Boston based company that creates “Business Solutions” through software. I work on the Relational database to make it more efficient by adding handwritten indexes to every java object…It is brutal and painstaking.
I started reading about the noSQL databases, like everybody, and one day I landed on the book “The Human Face of Big Data“, the case studies had epiphanic moments where the dots aligned (Jason Silva video) and the data mapped to the real world, it really inspired me to learn about noSQL technologies, and I started to appreciate the theoretical computer science courses that I had to take, for example, I learnt that Hash Trees are used by Cassandra.
Basically, the future is right now, we just might have found the key to unlock the limitless potential of human nature (ok, maybe…not?) — Netflix predicts House of Card’s Success.
THE PROJECT IN THREE WORDS: CROWD-SOURCED DEAL FINDING.
I am developing an Android app that allows users to post bargains or sales from stores. It would be like RedFlagDeals, but with location specific information.
Posting a deal:
If the user finds a deal he/she posts it to the app with a WHERE, WHAT and HOW MUCH. (Depending on the user’s location a short list of stores would automatically show up for selection). They are also encouraged to post a picture of the item or the price tag. Every post is location tagged for later retrieval and other users can rate, comment and flag the posts.
Finding a deal:
When the user wants to find a deal, depending on the user’s location a stream of items from nearby stores would be displayed in a newsfeed.
WHAT ROLE DOES CASSANDRA PLAY?
Cassandra is a column family database that has a very user friendly query language, CQL3, and scaling the analytics in future iterations of the app would be made much simpler by using an aggregate oriented database from the beginning.
For example, Kiipp (Scaling Riak at Kiip) had to jump over to Riak, which is like cassandra, to handle the massive amount of information points that they collect.
I am constantly learning and this could potentially lead to a Cassandra + Hadoop combination (How does Netflix predicts what you will watch?).
This is a 10,000 meter overview of the project, in the next post, I will highlight the technical and design decisions that went into the project.