Machines and data have been learning from quite a moment in time now. There is no domain left which does not use the stored data for analytic. Machines are getting better not only in terms of performance but also in terms of quality results because of the fact that they can learn. Learning is not an overnight process. It takes a sufficient and appreciable amount of time and effort to get trained, observe the patterns, record them and learn through them. Managing and measuring data have been two related mergers with data.
This definition of Data Science by Jennifer Widom is aptly put. She says big data is about two things: First is collecting large data and second is doing something with it. The process of going from first to second is the Data Science. This definition puts many of the current hot research areas together into a single basket.
Imagine a near future day where each and every activity of a person is tracked and monitored (or is it already?). Now, that is not an impossible task. A cell phone moving along in ones pocket can very well do it. Imagine what this data can do and what range of activities it can predict. Like you walk into a store and you get an alert of where to find your next purchase! Similarly, one can be alerted on where to find favorite food, where to catch the vehicle to commute and many such, all based on collected historical data.
The big data has impacted with larger benefits to domains like:
- Medicine
- Security
- Urban Planning
- Elections and
- Consumer Products to name a few.
What excites me the most is the e-commerce, online shopping and customer behavior data. In current scenario where people are opting to go for online money, there are many inventions unknowingly lined up. How do you see the online shopping domain progress down the line like in next 2-3 years? How customized should be your shopping experience?
With design and implementation in progress for my research related to the area, I would love to hear from you if you have something exciting or anything that you have come across which has given you a shareable experience (relating to online shopping and e-commerce). It could be a question, suggestion or a feedback. Let me know through comments or drop me a mail at prakash.hegade@gmail.com
PH Bytes 
Well that is something interesting you are doing research upon. The hottest topic of the century.
Well since there are already data science applied by many e commerce by studying their buying pattern, colour choosing to name a few. I feel in the near future, there will be alerts suggesting what to wear. The analytics can be applied to know ones wardrobe collection and prompt suggestions on what to wear for the day and for what occasion.
LikeLiked by 2 people
I do have included the color theory in my research, but not exactly this way. Will sure keep this inline. Definitely do-able.
Thank you so much Naveen for your thoughts.. 🙂
LikeLiked by 1 person
If I may reply to your thought Mr.Naveen, such system which suggests what to wear does exist, but that is by and large a combination of user inputs and user behavior. Color is definitely a good idea, albeit less thought one.
LikeLiked by 1 person
I got a chance to explore color theory in the regard. Understood a lot of things while implementing it that how new trends can be explored with rules coming from colors and art!
LikeLike
Amazon, which has announced a new grocery store that makes use of sensing and artificial intelligence to do away with the checkout altogether.
https://www.technologyreview.com/s/603035/amazons-grocery-store-doesnt-have-a-single-checkout/
I am not sure if we can relate this to e-commerce.
In the above article it is mentioned that Amazon keeps track of every item we purchase and adds it to list and charge us later.
It also removes items from the list we put back on the shelf…. Can this data(the items which he picked up but kept them back ) be used for suggesting him these items or recommending him similar items when there is some discount or offer.
LikeLiked by 1 person
There is a nice video for it on Youtube. Do watch it if possible. I did watch it just this morning:
Yes. It is going to be a new revolution in e-commerce and very much related to it. And it can bring in a lot of changes to recommendations systems.
Thank you so much for sharing your thoughts Rishabh. 🙂
LikeLike
Looks cool 🙂
LikeLiked by 1 person
That’s a very refreshing article sir. I think I have few thoughts on mind, please feel free to discard if they sound way too random.
** Relative recommendation
Currently the algorithms recommend for similar items (dress for dress, food for food). But I think its a smart requirement to have user behaviour not just studied but compared as well. Take it this way, suppose I buy Ready-to-eat meals quite frequently (with my location being recorded the same), the algorithm must be able to recommend- may be books(e-books rather) written on easy recipes for beginners, or nuts, shakes and healthy bars that can complement the packed food. If a person frequently purchases sporting products, the algorithm may well recommend Protein rich food and stuff related to sports but the point is the recommendation must be of products from “different domain”. (Currently available recommendation for different domains is quite naive- hair straightener is recommended if dryer is purchased and so on
Application must make use of its access to location and other details.
Say, having New Delhi as location, the algorithm must be able to recommend thermals in winters, sunscreen in summers and raincoats during monsoon. If the location is changed to say a snowing region, it may further recommend related necessities.
In the process, it must keep track of the location traversed. Say a person has stayed in location A for considerably a long time and moves to location B, the algorithm may suggest the person canned food of location A ( say tamarind chutney, pickles if he/she belongs to south).
Also, the algorithm may track the ordering behaviour of the customer. Some people tend to order then cancel, or add stuff to basket and discard due to shipping fees. There is never a shopping done in single click, it involves a plenty of iterations and reiterations, I have this strong feeling that this data can be closely studied to improve the purchasing probability ( Like suggesting a similar alternative which has no shipping fee, since the algorithm has pin code access I think it is doable) I somehow feel this is ine area where enough study has not been made. A little time on this can give us greater ideas I suppose.
There is this app called “health” (similar name. usually count the footsteps or record the running rate and distance) in almost all android phones. Since it is a built-in app I suppose the algorithm(e-commerce app) will have access to this data. If user is frequently monitoring his status implying he needs assistance in this area. Then recommendations like health food, sport wear, protein bars and so on may follow. Sounds nascent but sir I am sure you’ll build a concrete need for such stuff, if there is any.
If you meant to include e-learning systems in e-commerce, I have this recommendation. Most of the e-learning platforms are linked to the person’s email address. If the algorithm can learn by alerts set in mail account, or a commercial mail a person frequently gets. Say, I have subscribed to many websites regarding world economics, the e-learning platform may well recommend similar courses if they have any. This will have immense application in this era of competitive exams especially those who parallely work alongside preparation.
These are few things that came to my mind. I do not envisage greater benefit from these points they are quite raw and just touching the nuances, but I am sure your insights will make them constructive. I shall add in case anything more comes across my mind.
LikeLiked by 1 person
My sincere thanks for pulling out time and writing this. It really means a lot. Firstly, I cannot stop smiling as through my research, I am trying to put exactly what you have explained. Not just some static, predefined recommendations but smart, compatible and meaningful recommendations. The best part after reading your description is that now I can think of many horizons that I can expand my thinking and grow it to much bigger extent. It all makes so much sense.
I am done with around 20% of implementation and this is a motivation to pull inside much better things. I get it and it all holds good only to make that a better one. That order/cancel thing, had not thought earlier. It would be wise to incorporate that.
I will definitely think of an e-learning module. Had no thoughts thus far but now I see, it perfectly fits in.
Thank you so much. All this will help in getting a better quality. On it! Do let know if anything pops up in mind later ahead in time!
LikeLiked by 1 person
Very interesting stuff with a lot of possitieve points to the future
LikeLiked by 1 person
Thank you so much.. 🙂
LikeLike
Happy 2017
LikeLiked by 1 person
Thank you so much! 2017 Greetings!
LikeLike