Peer review by Shruti Sanghavi

[ShrutiSanghavi]

After reviewing this project, I feel that the team members were able to find a really interesting topic to work with. The importance of the project implementation was highlighted given the vast number of films released each year. Certain things that came to my mind include the following:

Things I like about the project:

1. The two problems posed by the team are very specific. Is successful, they will help estimate the result discretely. The team has assayed the varying factors and made an effort to even use tools of probability and statistics to help make more sense of the data. Furthermore, the market surrounding this problem is large, and this tool can greatly help not only big companies but also struggling artists to build strategies.
2. The dataset considers the time period of the release as a parameter, which I believe is an important aspect. For example, if a movie is released in a holiday period, people would be more free to watch the movie so it may actually do better than anticipated.
3. With the growing social media revolution, an inclusion of the social media statistic can gravely enhance the results of the predictions to the problems posed. This is a very real measure of the market and will add an honest dimension to the analysis.

Uncertainties about the project:

1. The two problems posed by the team will have to be treated independent, as very frequently, there is no relation between the IMDB rating and the performance at the box office. So either they are addressed separately or one is chosen and worked on.
2. Although including the social media statistic will add an interesting aspect, should Facebook shouldn’t be the only one to consider? Will that be appropriate? Or YouTube be a better estimator?
3. Although movies are segregated by genre, the emotional aspect and relatability of the content with viewers is an important factor. Usually movies inclining to patriotism or conspiracies have a greater impact on the audience than others.
4. The IMDB rating is based on the votes by the regular voters. Is this data biased by any chance?