The objective of the project is to predict a movie's total gross based on information about the movie that can be collected in the first week following its release.
I like how the review and initial gross data is restricted to the first week, because that means that the model can be applied to predict the gross of future movies. The data visualizations show some clear trends (especially for the non-review based features), which will give good insight going forward for feature and model selection.
One point of concern is that based on your 3-d visualization the metascore does not appear to be a strong predictor for the movie's financial success. Also, regarding the formatting of the feature descriptions in this report, in the future it would be much clearer if you used a table - the current formatting is a bit difficult to read. The trend for open-theaters does not appear to be linear, so perhaps in the future a non-linear model could be used.
A potentially interesting feature could be search frequency or number social media mentions of the movie, since that might be a better indicator of popularity than reviews. Of course, this would require additional data scraping but it could be interesting to look into.
The objective of the project is to predict a movie's total gross based on information about the movie that can be collected in the first week following its release.
I like how the review and initial gross data is restricted to the first week, because that means that the model can be applied to predict the gross of future movies. The data visualizations show some clear trends (especially for the non-review based features), which will give good insight going forward for feature and model selection.
One point of concern is that based on your 3-d visualization the metascore does not appear to be a strong predictor for the movie's financial success. Also, regarding the formatting of the feature descriptions in this report, in the future it would be much clearer if you used a table - the current formatting is a bit difficult to read. The trend for open-theaters does not appear to be linear, so perhaps in the future a non-linear model could be used.
A potentially interesting feature could be search frequency or number social media mentions of the movie, since that might be a better indicator of popularity than reviews. Of course, this would require additional data scraping but it could be interesting to look into.