Group Case Study #$1($Chapter $6)$

Big Data: The Case of Google Flu Trends

Question #$1$: In relation to Google$$s GFT and Netflix$$s viewing recommendations, I believe that Netflix is a predictive system. This is mainly due to Netflix collecting data on which shows and movies you have previously watched, what genres$/$categories they fall into, and utilizing this information in order to provide a curated list of similar shows and movies based on this data collection. Therefore, this is directly predicting your watch behavior.

Similarly, Google$$s GFT is also a predictive system, which operates in order to provide predictions based upon recent search queries. While in contrast to Netflix$$s data collection history, this isn$$t as cut and clear as what show you watched. With individuals typing in questions and searches into google, this can make the system $$quite vulnerable to overfitting to seasonal terms unrelated to the flu$($Lazer & Kennedy, $2015).$ In this, random searches containing certain keywords can greatly offset the prediction estimate, leading to the premise that I believe Netflix$$s system is more straightforward in their predictions.

Question #$2$: The CRISP$-$DM data mining process can be described as a $$structured approach for the planning and execution of a data mining project$$ through the means of six stages $($Stair et al$.,2020).$ In relation to the development of Google Flu Trends, I would say that the data understanding aspect of the CRISP$-$DM model definitely relates together. In this, Google GFT would utilize data understanding to highlight certain trending searches with the goal to come to a prepared conclusion on whether influenza was highly present. This same idea applies to Google$$s updates to the system through their adjustments of keywords that would trigger a flu trend. With this, the aspects of business understanding, data understanding, and data preparation greatly relate to the tasks of understanding a goal, compiling data, and learning from this data in order to evaluate a prediction that would deploy to the public in relation to influenza presence.

Reply to the discussion post.