Applications on Big Data Analytics



so now that we have studied what are the most important components of big data analytics it is important to see what are the real world use cases so here are a few of those use cases of course there are many many but here a few of them so like telemetry data analysis telemetry data is any kind of data which is collected based on the system behavior and is generally collected in a passive manner so for example Windows machines might be collecting what processes are running it at any point of time electricity boards might be collecting your electricity consumption and so on so analyzing that data by a propensity model so what is the likelihood that this guy who is just visited your store will actually buy the stuff buy the product right social network analysis predictive maintenance predictive maintenance is actually important across a large number of industries a large number of lifts or elevators run all across cities before our elevator goes down or you know some accident happens you want to really predict before the actual accident happens and do the maintenance beforehand usually rather than any predictive maintenance what many lift companies do is to do a maintenance at a particular time periods now this may not be very nice if a particular lift is being used at a much higher load and therefore doing predictive maintenance is very important but if your maintenance is also important another other kind of critical equipment for example medical refrigerators are also like parts of an aeroplane so all of these of course can undergo periodic maintenance but sometimes they can actually go bad very quickly and therefore it is important to do such predictions as to when this machine can go bad and therefore do a maintenance beforehand so there are many other use cases like weather forecasting which actually is very important for Indian farmers most of Indian farmers depend on rain fed irrigation basically saying that you know if there is gold rain they actually have very good yield so therefore it is very important for them to know the weather forecasts so that they can sow the seeds at the right time healthcare outcomes for patients and hospitals churn analysis again for a whole bunch of companies who want to know how what percent of their users and actually which users will actually move out from their platform and move on to some competitive platforms right fraud detection life sciences research and so on they're whole bunch of use cases and it actually turns out that these use cases also also exist across a large number of domains and listed a few of these domains here now some of the interesting use cases are in banking and insurance insurance actually has multiple interesting problems when when you apply for an insurance the insurance company somehow has to figure out what should be the premium value if the premium value is fixed as too high then folks don't want to buy your insurance but if the insurance value is fixed to low the premium value in that case if something wrong happens with the insurer then the insurance company ends up paying too much so the optimized premium value is something that can be predicted using machine learning that said there are of course more use cases in insurance for example there's a whole bunch of fraud that exists around insurance especially around health insurance people get sick and they visit hospitals hospitals when they know that you are insured sometimes can actually do a whole lot of fraud for example they might do some certain procedures but may charge much more right charge the issuance agency much more so in those cases using machine learning so as to figure out whether there is a fraud or not in the insurance claim is also important machine learning problem right so similarly there exist many many use cases and interesting use case comes to mind in the fashion industry now you would think what fashion has to do with any science or machine learning it turns out there are very interesting use cases recently an has been developed using machine learning which can actually take a photograph of your closet and then depending on the occasion it can actually suggest what kind of clothes combination of clothes makes sense for you so essentially it's an outfit suggestion app and you know where do you get training data for learning such app well there are so many photos on the web which talk about what things go with what and somebody could really label those things as as cloth entities and therefore use that training data to train a model which given an occasion recommend the best outfit so overall to sort of summarize machine learning we looked at various ways of doing machine learning data science and so on and it's so important now is actually prevalent across all domains

2 Comments

  1. Karthikeyan S said:

    plzzz explain

    June 26, 2019
    Reply
  2. Karthikeyan S said:

    is data science and big data is same

    June 26, 2019
    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *