Slides Framework

  • Casting Actors for Your Movie

    Casting is one of the most crucial parts of the filmmaking process; choosing the actors can make or break a film. Choosing the right actors for a movie is important because they are the ones who bring the characters to life and bring the script to fruition. When the right actors are chosen for a movie, it can significantly enhance the overall quality and enjoyment of the film. On the other hand, if the wrong actors are chosen, it can have a negative impact on the film and may result in poor reviews and box office performance. Therefore, it is important for filmmakers to put a lot of thought and consideration into casting the right actors for their movies.

    Let's start the journey!
  • cinema

This project aims to predict how successful an specific actor would be if they were cast in a specific role, and to rank the actors by the most appropriate for the role to the least appropriate, which would serve as casting suggestions to the directors and producers.

  • 1- Does hiring the best actors guarantee a successful movie?

  • 2- Do the best actors play the same roles over and over again?

  • 3- Does hiring an actor who played the same role over and over guarantee a successful movie?

  • 4- Who are the most appropriate actors for your movie?

First off

Does hiring the best actors guarantee a successful movie?

We define a successful actor as one who has been awarded an Oscar by the Academy of Motion Picture Arts and Sciences, since this reflects the opinions of other industry professionals.
We consider a movie to be successful if it has high ratings from the general audience, which we will get from the IMDb database.
We will examine the relationship between having a successful actor and the success of a movie in this part.

  • Our actors' dataset is composed of

    numbers

    actors

  • Our movies' dataset is composed of

    numbers

    movies

    In order to explain such a relationship, we do a matched analysis comparing the actors who have won an Oscar, with actors that have not been nominated. To make matches more comparable, we consider pairs who have the same movie release era, movie languages, and movie countries. After doing a 1-1 matching, we use linear regression model to estimate the effect of having won an Oscar in the final movie rating, which we find to be positive and quite significant.

Secondly

Do the best actors play the same roles over and over again?

We define once again a successful actor to be one that has been awarded an Oscar by the Academy of Motion Picture Arts and Sciences.
We find the roles that actors play using the plot summaries; after running the plot summary through the Stanford NLP Core Pipeline, we extract specific linguistic features for each character which are then clustered into different personas.
We will examine the relationship between having played the same persona a lot of times and having won an Oscar award.

  • Generated bag of words for every character in movie plot

    Choose a movie

    • Ooruku Nooruper
    • The Assassin
    • El diputado
    • Breakthrough

    • Plot :


      The story follows Balan , a young artist who becomes disillusioned by the current political situation of his society. He joins the revolutionary organization Ooruku Nooruper, accidentally kills a priest and is sentenced to death. The film examines various issues related to capital punishment.

      See the bag of words!

    • Plot :


      British private detective Edward Mercer is employed to travel to Venice and locate an Italian who is to be rewarded for his assistance to Allied airmen during the Second World War. Once he arrives in Italy, however, he becomes mixed up in an assassination and a great deal of mystery.

      See the bag of words!

    • Plot :


      Madrid, Roberto Orbea is a member of the Spanish communist party. He is married to Carmen , and he has been elected as Deputy in the first democratic elections in Spain. But his enemies, the fascist, know his double life. Roberto likes boys, and they hire Juanito to seduce the politician. They fall in love…

      See the bag of words!

    • Plot :


      Starting in late May 1944, during the German retreat on the Eastern Front, Captain Stransky orders Sergeant Steiner to blow up a railway tunnel to prevent Russian forces from using it. Steiner's platoon fails in its mission by coming up against a Russian tank. Steiner then takes a furlough to Paris just as the Allies launch their invasion of Normandy. Steiner's unit is transferred to France, occupying the village of St. Bologne. General Hoffman orders Steiner to cross into enemy territory and confer with American Colonel Rogers and General Webster that the High Command of the German Army is plotting to assassinate Hitler and would like to surrender. The plan fails and American forces launch an attack on German forces in St. Bologne where Stransky has planned an explosion to destroy both the Americans and civilian inhabitants.

      See the bag of words!

  • Bag of Words for Cluster 22

    Choose a character name

    • Irene
    • Hitler
    • Ryan
    • Caparzo
    • Wade
    • Ryan
    • Bootsmann
    • Hinrich
      • attribute : die
      • patient : assassinate
      • attribute : be
      • attribute : fourth
      • attribute : francis
      • attribute : ryan
      • attribute : feel
      • attribute : salute
      • attribute : miss
      • patient : send
      • attribute : class
      • attribute : stand
      • attribute : private
      • attribute : private
      • attribute : lead
      • attribute : frederick
      • attribute : say
      • patient : find
      • patient : tell
      • attribute : james
      • attribute : distressed
      • attribute : ask
      • patient : visit
      • patient : die
      • attribute : defend
      • patient : elderly
      • patient : american
      • patient : bring
      • patient : locate
      • patient : earn
      • patient : first
      • attribute : bleed
      • patient : wound
      • attribute : die
      • attribute : wade
      • patient : wound
      • attribute : dangerous
      • attribute : ryan
      • patient : wound
      • patient : wound

    We do a matched analysis comparing the actors who have won an Oscar, with actors that have not been nominated. We use linear regression model to estimate the effect of having won an Oscar in the ratio of the total roles played that have the same persona as their maximum occurrence persona, which we find to be positive and quite significant; an Oscar award winner is plays on average two times as much their maximum occurrence persona.

    We investigate the effect of having played the same role before directly in the movie rating. We do a matched analysis comparing the actors who have played the role before, with actors that have not. We use linear regression model to estimate the effect of having played the role before in the final movie rating. The results of this analysis are inconclusive.

Finally

Who are the most appropriate actors for your movie?

In order to predict how successful picking a specific actor for your movie is, we combine the features previously discussed (Oscar wins or nominations, list of personas played) with the actors’ metadata, and we train a model to predict the final movie rating.
To pick which features should or shouldn’t go into the model, we use forward feature selection; we start with a model that doesn’t include any feature, and we continue adding feature one by one until the adjusted R-squared metric doesn’t improve anymore. At every step, we add the feature with which the adjusted R-squared metric is maximized. Adjusted R-squared is a measure of the goodness of fit of a statistical model that has been adjusted for the number of predictors in the model. It increases when a new predictor significantly improves the model, and decreases when the improvement is less than expected. It can be used to compare the performance of different models and determine which model provides the best fit to the data.

The model can then be used by specifying the persona of the role we want to cast for, as well as any important metadata (e.g. gender or ethnicity). We filter the list of actors by the necessary metadata constraints and calculate the relevant additional features that comes from the persona of the role (e.g. the number of previous roles the actor has played that fir into the relevant persona). We then predict the final movie rating for each actor that fits into the metadata constraints using the model previously mentioned.

  • Suggested actors for the role

    Choose a character

    • Shutter Island - Rachel Solando
    • Shutter Island - Dr. John Cawley
    • Shutter Island - George Noyce
    • Shutter Island - Teddy Daniels
    • Shutter Island - Dolores Chanal
    • Shutter Island - Chuck Aule
    • Shutter Island - Andrew Laeddis
    • Pirates of the Caribbean: On Stranger Tides - Captain Teague
    • Pirates of the Caribbean: On Stranger Tides - Blackbeard
    • Pirates of the Caribbean: On Stranger Tides - Hector Barbossa
    • Pirates of the Caribbean: On Stranger Tides - Angelica Teach
    • Pirates of the Caribbean: On Stranger Tides - Syrena
    • Pirates of the Caribbean: On Stranger Tides - King George II
    • Pirates of the Caribbean: On Stranger Tides - Joshamee Gibbs
    • Pirates of the Caribbean: On Stranger Tides - Philip
    • Superman III - Ross Webster
    • Superman III - Vera
    • Superman III - Perry White
    • Superman III - Superman
    • Superman III - Lois Lane

Thank you for your attention!

  • Generated bag of words for every character in movie plot

    Choose a character name

    • Nooruper
    • Balan
    • Go back to plots!
      • attribute : ooruku
      • attribute : organization
      • attribute : revolutionary
      • patient : join
      • attribute : young
      • attribute : join
      • attribute : kill
      • patient : sentence
      • attribute : balan
      • patient : follow

  • Generated bag of words for every character in movie plot

    Choose a character name

    • Mercer
    • Go back to plots!
      • attribute : british
      • attribute : private
      • attribute : mixed
      • attribute : detective
      • patient : employ
      • attribute : edward
      • attribute : arrive
      • attribute : young
      • attribute : join
      • attribute : kill
      • patient : sentence
      • attribute : balan
      • patient : follow

  • Generated bag of words for every character in movie plot

    Choose a character name

    • Juanito
    • Orbea
    • Go back to plots!
      • attribute : seduce
      • name: Orbea
      • patient : elect
      • attribute : like
      • attribute : madrid
      • attribute : roberto
      • attribute : orbea
      • patient : marry

  • Generated bag of words for every character in movie plot

    Choose a character name

    • Hitler
    • Rogers
    • Webster
    • Stransky
    • Steiner
    • Go back to plots!
      • patient : assassinate
      • attribute : american
      • attribute : colonel
      • attribute : general
      • attribute : plan
      • attribute : sergeant
      • attribute : confer
      • attribute : take
      • attribute : blow
      • attribute : cross