Assignment 1: Data Analytics ModelShort paper: 15% of Final GradeThe purpose of your first Assignment in our course is to identify an area of interest of yours where large or big data analytics might be an applicable method to discover new relationships among variables or data components. Be very reasonable while selecting this area, as this assignment is a precursor to your subsequent assignments. [It is recommended that you pick up a topic where you can get publicly available data and can run simple analysis using R- but not this week, this is if you use the same topic for the last assignment.]This paper has three parts to it:1. Identify a big data area applicable to analytics;2. Specify or suggest how data in this area might be identified or explored;and,3. Articulate a model and possibly a method for such exploration. Your project here is hypothetical at this point so the paper is only proposing a study.The purpose of this assignment is to create a blueprint for other work you do in this course.HOW – TO – DO – THIS – ASSIGNMENTHere is how you might approach this assignment. You can start with ideas you have already offered in our Discussion Boards or you can begin de novo with a new idea. You should then do a search of pertinent literature, reports, or research on your topic (or on a topic close to yours which could serve as an exemplary example). That search can be done in Google, Google Scholar, or in a library database. I will also try to pepper our Doc Sharing part of eCompanion with more reports.Be sure to cite any work you use in an appropriate way (see postscript at end). Here is an overly detailed explanation to guide. The content of your paper is important even though this assignment is only worth 15% of your grade. Find out what big data efforts have been constructed in the area of your interest. Build from their models. Remember to focus on their model or plan, specify the way in which they did analytics, and give a brief summary of their findings. Try not to ignore these steps (also articulated in the next paragraph). If there is no big data effort in your area of interest, then you will need to explore a similar area and build from that.Here are the three arbitrary steps you might want to consider if you are not grasping the overall effort here.1.1.Introduce your topic and give some background information.2.It would be good if you could give a brief example of what the data set(s) looked like in your area or in an analogous area that you are using as a model AND what it might look like in your hypothetical project. The data set(s) could be put in a table or a mini-spreadsheet. If you are planning to use an existing dataset, then please specify.3.You should try to display the model for the similar study and the model for your study. Those two models could be very similar or almost identical and that could be appropriate. Usually these models are depicted as a figure with boxes and arrows etc.4.Indicate how you might analyze your “data” by:a.Specifying which relationships might be explored; andb.What you might expect to find from such explorations. This last part is not a hypothesis but a speculative, informed estimate.5.Discuss what might have to be done to have your plan/model actually get set-up for analysis, if you are imagining some new kind of data collection.Your paper presents a model and a possible strategy for analysis.LENGTH of the ASSIGNMENT: I would think that an individual effort might be several pages (ie 2-4; don’t care about spacing). How your idea unfolds and your plan of action is more important than the length. Citation style should be simple, accurate, and consistent.Examples.You want to predict population growth in NJ towns. You gather data from the US Census and see if you can find a mathematical fit to the growth of towns, perhaps just an assumption that population is linear in time. You test this by trying to predict 2010 from the previous decades. You would then think you can say where new schools or other public facilities would have to be built. You would like to have an explanation – perhaps growth is related to distance from NYC or Philadelphia; perhaps to average temperature (unlikely); perhaps to distance from the Turnpike. So you test for correlations and see if anything seems important.You find a newspaper article about a computer that writes pop music songs (an example on Youtube is Daddy’s Car from Sony labs. I am not recommending it). You decide you can come up with something that will write imitation Harry Potter stories. You collect several thousand fanfiction Harry Potter short stories, find the most frequent words in them, and see what phrases they occur in and string those phrases together. (Something like this has been used to write computer science papers). You see whether the stories are readable and whether you can find a way to make them more readable by seeing when phrases go together.You want to find new drugs to treat some disease. You find a newspaper article about the AstroZeneca (giant pharmaceutical company) challenge, in which they are publishing properties of some candidate chemical molecules and asking individuals to evaluate their structures as potentially useful for treating cancer. You think of doing that for some other disease – find relevant chemicals, see what their properties are, see what the properties of known useful drugs are, and try to identify the chemicals that might be useful.
