Revista Latina de Comunicación Social. ISSN 1138-5820
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial-SinDerivadas 4.0
Emotional Design in Fiction Series: Sentiment Analysis of the Script and Audience's Emotional Response
Enrique Guerrero-Pérez
Universidad de Navarra. Spain.
Patricia Diego
Universidad de Navarra. Spain.
Research project “Quality as an essential factor for a sustainable business model in streaming platforms”, funded by the Ministry of Science, Innovation and Universities of the Government of Spain. Start date: 09/01/2024. End date: 08/31/2027. Reference PID2023-150258NB-I00.
How to cite this article / Standard reference:
Guerrero-Pérez, Enrique, & Diego, Patricia (2027). Emotional Design in Fiction Series: Sentiment Analysis of the Script and Audience's Emotional Response. Revista Latina de Comunicación Social, 85, 1-21. https://www.doi.org/10.4185/RLCS-2027-2613
Date of Receipt: Oct. 3, 2025
Date of Acceptance: Nov. 12, 2025
Anticipated Publication Date: Jan. 19, 2026
Date of Publication: Jan. 1st, 2027
ABSTRACT
Introduction: This article explores the use of sentiment analysis in scriptwriting for fictional series, with the objective of creating content that resonates with the audience, thereby enhancing the quality and sustainability of the final product. The article is based on the hypothesis that emotional gratification optimizes audience loyalty by appropriately developing plots and characters. Methodology: A case study of the Spanish thriller Ana. All in (Ana Tramel. El juego) is conducted by triangulating three analyses: one of the pilot script using R, one of Twitter messages during the television premiere, and one of a questionnaire administered to a sample audience after a private screening of the pilot episode. Results: The emotional design of the script is effectively conveyed (the main emotions—trust, sadness, and fear—are consistent), yet the audience unconsciously reinterprets it, emphasizing fear. Social media reactions are more polarized and negative. However, conscious analysis highlights surprise and anticipation as key to engagement. It also validates the effectiveness of complex characters who generate positive feelings through trust and empathetic sadness. Discussion: Comparing the screenwriting and production teams' intentions with the audience's unconscious reactions on social media and conscious evaluations confirms these studies' suitability as complementary tools in creative processes, aligning narrative design with audience emotional reception. Conclusions: This methodological triangulation is appropriate for understanding the viewer experience. The emotional design of characters and plots is fundamental to achieving audience engagement and offers guidelines for optimizing the quality and sustainability of audiovisual production.
Keywords: television, production, fiction series, audiences, emotions, feelings, script, engagement, sustainability.
1. INTRODUCTION
The audiovisual industry is at a turning point driven by a technological revolution that is redefining how content is produced, distributed, and consumed (Quintas-Froufe & González-Neira, 2021; Doyle, 2016a; Francés i Domènec, 2015). This study examines how applying big data techniques, such as sentiment analysis of scripts and audience feedback, can contribute to producing quality content that engages audiences in an increasingly competitive environment.
Digital convergence has led to the merger of three industries: media and entertainment, information and communication technologies (ICTs), and telecommunications (López Villanueva, 2011; Vizjak & Ringslstetter, 2003). Consequently, the current audiovisual model relies on connectivity and interactivity (Holt & Sanson, 2014) as well as an attention economy fueled by access to an extensive catalog of content (Vounderau, 2015).
This new production and distribution paradigm poses significant economic, ecological, and social sustainability challenges (Doyle, 2018). The project development phase is crucial, so it is essential to design high-quality content that adapts to the interests and sensibilities of the audience, especially young people. Their audiovisual behavior is more susceptible to technological changes and the convergence of television and the internet (Guerrero, 2018; Saló et al., 2022).
Traditional audience measurement methods significantly limit the ability to accurately reflect audience behavior in this multiplatform digital environment (Rodríguez Vázquez et al., 2019). This has led to the search for new indicators, such as engagement. Engagement is a multidimensional concept involving involvement, connection, and experience. It is also related to motivations, as well as emotional and cognitive responses (González-Bernal, 2016).
The engagement process has three phases: before, during, and after. The first phase is the expectations phase. The second phase is the thoughts, emotions, and consumption practices phase. The third phase is the subsequent behaviors, such as sharing or saving content to a personalized playlist, phase (González-Bernal, 2016). Heath (2007) further explores this concept, defining it as a subconscious emotional construct dependent on the intensity of feelings generated when processing a message.
However, the factors influencing perceived quality and engagement are complex due to their subjective nature (Etayo et al., 2023; Kimber-Camussetti & Guerrero-Pérez, 2022). Social media plays a significant role in strengthening the audience's emotional connection with content (Atarama-Rojas & Feijoo, 2023; Fernández-Gómez & Martín-Quevedo, 2018; Castro-Mariño, 2016).
The Uses and Gratifications Theory (Katz et al., 1973) provides a solid conceptual framework for this research in relation to the aforementioned challenges. Despite technological innovations, the fundamental needs that audiovisual entertainment media fulfill have remained constant (Cuesta Cambra et al., 2021). The theory explains how and why audiences use media, emphasizing emotional needs and offering a valuable perspective amid increasing supply and digital convergence (Jenkins, 2006).
Bartsch (2012) delves into the crucial role of emotions in audiovisual experiences, identifying amusement, tension, and empathic sadness as key factors in emotional gratification. Notably, even emotions stemming from sad or tragic events can offer gratification through controlled arousal, as described by Zillmann (1988, 1996) in his Mood Management Theory. Furthermore, Busselle and Bilandzic (2009) argue that audiences can experience emotional gratification through identification with characters, which allows them to have vicarious experiences.
In the face of these challenges, the audiovisual industry is turning to new scientific techniques based on data analysis and computational science. These techniques aim to improve creative processes and adapt content to audience sensibilities. These tools can help showrunners and screenwriters design plots and characters that evoke empathy and maintain audience attention in an oversaturated market (Higueras-Ruiz et al., 2021). For example, recent studies have analyzed the application of neuroscientific methods to fiction series and audiovisual advertising production, confirming their effectiveness in fostering audience loyalty and enhancing engagement (Guerrero-Pérez et al., 2025; Tapia Frade et al., 2025).
Specifically, sentiment analysis applied to scripts is emerging as a promising technique for producing quality, audience-tailored content, improving engagement, and facilitating more sustainable production in place of oversized catalogs. However, using these techniques also presents limitations and risks to creative freedom that must be considered, as well as potential biases.
In short, understanding the emotional design of scripts and the audience's emotional response can help produce quality content that fosters engagement and sustainability. Nevertheless, measuring the suitability of narratives and characters to audience sensibilities and the quality of audience attention remains a significant challenge for current audience measurement systems. Applying scientific data analysis methods to creative processes, such as sentiment analysis of scripts, offers new perspectives for addressing these challenges.
2. OBJECTIVES
This article explores the relationship between the emotional design of fictional series scripts and audience response. The purpose is to enhance engagement through emotional gratification and create content that is more relevant to target audiences. This purpose is divided into three objectives.
The first objective is to identify emotional patterns in the scripts using sentiment analysis techniques with R software. This type of analysis detects emotions implicit in the plots and characters and identifies whether the emotion is positive or negative. This helps understand how screenwriters can imbue scripts with emotional intentions to make them more appealing to the audience.
The second objective is to understand the audience's conscious and unconscious emotional responses to the series in order to promote emotional gratification and adapt the content to their psychological needs.
The third objective compares the emotions present in the script with the audience's emotional response. This paper explores the relationship between emotional content design and emotional gratification to foster engagement. The paper is based on the hypothesis that appropriate emotional gratification increases viewer loyalty and a series' likelihood of success, thereby contributing to the sustainability and quality of the production model.
3. METHODOLOGY
This study analyzes the pilot episode of Ana. All In, which is a series produced by Televisión Española (TVE). Titled "La apuesta" (The Bet), the 56-minute episode aired in primetime on La 1 on September 21, 2021. According to Kantar Media data (GECA, 2021), it achieved a 9.3% audience share and an average viewership of 1,259,000. It is currently available to stream on Netflix and RTVE Play, TVE's free streaming platform.
It was selected as the most representative episode of the series. It introduces the main storylines and characters, showcasing the production's narrative potential.
Ana. All In is a television adaptation of Roberto Santiago's novel of the same name (Ana Tramel. El juego), with Santiago and Ángela Armero writing the script. The series is a co-production of RTVE, Tornasol, ZDF, and De A Planeta with an approximate budget of €5 million.
The plot revolves around Ana Tramel, a highly prestigious criminal lawyer going through a profound personal and professional crisis marked by dependence on drugs and alcohol. Relegated to handling minor cases at the firm of her friend and partner Concha, who is also dealing with a conflict stemming from domestic violence, Ana's life takes an unexpected turn.
The narrative turns when her brother, Alejandro Tramel, is arrested and accused of murdering the director of the Gran Casino de Robredo. His unexpected plea for help forces Ana to become involved in the case. In the process, she meets Lieutenant Moncada, the investigator, and begins to develop feelings for him. She discovers her brother's hidden life, which includes his partner, Helena, a young Russian stripper; a two-year-old son; a serious gambling addiction; and substantial debts.
This confluence of circumstances compels her to accept the defense, marking her return to high-level legal practice after five years away from the courts. It also marks her direct confrontation with the powerful interests of the gambling industry. For the purposes of this study, it is important to note the role of Lieutenant Moncada. While he is portrayed as a reliable authority figure in this pilot episode, subsequent episodes reveal him to be the main antagonist of the plot. His relationship with the protagonist shifts from attraction to confrontation.
This series was chosen for several reasons. First, it is a police thriller that delves into gambling addiction, a socially relevant topic. Second, it contains storylines with a strong dramatic component, such as gambling addiction, alcoholism, and murder. This is essential for analyzing the emotional design of audiovisual content and its capacity to build audience loyalty and engagement. Third, it possesses qualities that distinguish it from other titles: it aired during primetime on Spain's main public television channel, and it has been part of the Netflix catalog since then. The production involved renowned figures such as Gerardo Herrero, an award-winning producer with an extensive career, and actresses Maribel Verdú and Natalia Verbeke. Furthermore, when the study began, the series had just premiered, which facilitated access to original production materials, such as the script.
The research applies quantitative and qualitative methodologies. Thus, it combines text processing techniques, such as sentiment analysis, applied to the script and audience messages on social media with questionnaires administered to a sample of viewers to allow for the study of their conscious emotional reactions.
3.1. Sentiment Analysis in R
To analyze the emotional design of the script and the audience's emotional responses, natural language processing (NLP) and text mining techniques were employed using R (version 4.4.3) via the RStudio software. Specifically, sentiment analysis was applied to the texts using the Syuzhet package (version 1.0.7) to identify emotional shifts related to conflict in the narrative structure. This tool extracts sentiment and narrative arcs using various sentiment dictionaries (Jockers, 2015). In this study, the NRC Manual Lexicon (Mohammad & Turney, 2013) was used, given its versatility and applicability to all types of texts. In this study, it was applied to both the pilot script and audience messages on social media.
However, as with any methodology, this one has limitations:
- Literalness: The automated application of this type of lexicon in sentiment analysis has difficulty interpreting double meanings, irony, sarcasm, double negatives, and other rhetorical devices that alter the literal meaning of the text (Kiritchenko et al., 2014).
- Language: Most lexicons (dictionaries) for sentiment analysis using large volumes of data are optimized for English. Analyses in other languages lose precision. Although the NRC lexicon can be used in Spanish via machine translation, it lacks cultural nuances.
- Volume: The robustness and comprehensibility of the results of this type of analysis improve significantly with larger volumes of processed textual data (big data).
Despite its limitations, sentiment analysis is useful for understanding the emotional nature of any text. Specifically, it allows for the investigation of polarity (whether positive or negative) and the emotion evoked (Mohammad & Turner, 2013). Emotions themselves are neither positive nor negative, but the feelings they generate are, depending on whether they are favorable. Emotions are instinctive and automatic responses to a stimulus, while feelings arise from cognitive reflection on that stimulus and the individual's response (Nünning, 2017).
The NRC lexicon can detect eight basic emotions, which are universal and present in all cultures. These emotions were identified by Plutchik (1962, 1980, 1994) and classified into four pairs of opposites: joy and sadness, anger and fear, trust and disgust, and anticipation and surprise.
3.1.1. Analysis of the Audiovisual Script
The final version of the episode's script was used for the sentiment analysis. Consisting of 55 pages and 47 scenes (reduced to 45 after the pilot episode was produced), the final version totaled 12,106 words distributed across 1,369 lines. To process the text with the NRC lexicon, punctuation, accents, numbers, and Spanish stop words (Multilingual Stopwords List 2.3) were removed. The text was then tokenized using sentences as the unit of analysis.
With this technique, only the script was analyzed, not the entire audiovisual production. Thus, the results derived from this part of the methodology are obtained exclusively from the text's sentiment analysis and do not consider acting, directing, production, music, cinematography, and so on, all of which significantly influence the emotions perceived by viewers (Nünning, 2017). Furthermore, even though the latest version of the script was used, the script was modified during filming and subsequent editing.
3.1.2. Analysis of Audience Reactions on Social Media
In addition to studying the script, a sentiment analysis was applied to messages posted on social network X (formerly Twitter) during the live broadcast of the September 2021 episode on La 1 of TVE, from 10:44:34 p.m. to 11:40:06 p.m. UTC+2, to examine the audience's unconscious emotional response. The tweets were extracted in real time using the Twitter Developer API and the R libraries twitterR (1.1.9) and rtweet (2.0.0). Only Spanish messages that used the official television premiere hashtag, #EstrenoAnaTramel, were selected. After filtering out retweets, a final sample of 197 messages containing a total of 3,006 words was compiled.
The processing of these tweets followed the same protocol as the script analysis: text cleaning to remove non-alphanumeric characters (URLs, emoticons, accents, symbols, etc.), Spanish stop words (Multilingual Stop Words List 2.3), and phrase-level tokenization. Subsequently, a sentiment analysis was performed using the NRC manual lexicon from the Syuzhet package.
This computational analysis methodology has been extensively tested in other research projects studying audience reactions on social media and in other contexts from an emotional perspective. Examples include the perception of television content quality (Amalia et al., 2018), analysis of online film reviews (Bedi & Tamrakar, 2019), success of content on streaming platforms (Malik et al., 2022), political pluralism in talk shows (Ceron & Splendore, 2018), conversations among online gamers (Ángeles Gómez & Quintana López, 2019), the impact of the pandemic on social media (Cebral-Loureda & Sued-Palmeiro, 2021), and its representation in serialized fiction (Chicharro-Merayo et al., 2022).
Taken together, these studies confirm the validity of the sentiment analysis methodology for understanding audience perceptions across various fields. Thus, this technology is establishing itself as a valuable tool for evaluating quality, anticipating success, understanding different sensitivities, and strengthening emotional connections to audiovisual content.
3.2. Questionnaires for an Audience Sample
To complement the sentiment analysis, a private screening of the pilot episode was conducted with 59 people divided into two groups in November 2021. Conscious emotions related to the series and its characters were assessed using an anonymous questionnaire. This technique enables comparison of emotions derived from sentiment analysis of both the script and messages posted on X, providing a more comprehensive perspective.
The sample included 24 men (40.7%) and 35 women (59.3%) between the ages of 18 and 30. All participants were enrolled in or had completed university studies, and Spanish was their native language. The study focused on the youngest segment of the adult audience because this demographic exhibits disruptive and interesting audiovisual consumption patterns that will influence the future of the industry (Guerrero-Pérez, 2018; Saló et al., 2022). Additionally, the sample was slightly female-dominated to align with the audience profile of the episode's television premiere (44.5% men and 55.5% women, according to GECA, 2021).
Immediately after viewing the episode in the same screening room, participants completed an anonymous online questionnaire. The questionnaire was designed to conduct a descriptive analysis of the participants' viewing habits and to explore the emotions evoked by the episode and each of the main characters. Respondents also indicated whether these emotions translated into positive or negative feelings. In addition to the eight basic emotions identified by Plutchik (1962; 1980; 1994) and included in sentiment analysis in R using the Syuzhet package (NRC lexicon), "contempt," as defined by Ekman (1972; 2007), was added for the identification of emotions in this part of the research. Thus, the questionnaire asked the audience sample about nine emotions in total, providing a more comprehensive analysis.
The questionnaire is divided into the following sections (see Annex):
Section 1: Individual identifier (numbered wristband to anonymize responses).
Section 2: Demographic variables (gender and age).
Section 3: Audiovisual habits (media and device preferences, most frequently consumed content, viewing time, whether viewing is done alone or in a group, etc.).
Section 4: Emotions and feelings evoked by the episode and main characters.
Integrating these methodologies—automated sentiment analysis of texts and analysis of conscious emotions through questionnaires—is a unique aspect of the research. This multifaceted, complementary approach allows for a better understanding of the success or failure of fictional narratives through emotional engagement, shedding light on the complex phenomenon of emotional content design and audience emotional response. This knowledge can then be used to improve the quality and sustainability of production processes.
4. RESULTS
The data analysis is presented in three sections that correspond to the research objectives. First is the emotional design of the script. Second is the audience's emotional response, as expressed through social media posts and a questionnaire. Finally, there is a comparison between the emotional design and the audience's conscious and unconscious responses.
4.1. The Emotional Design of the Script
Sentiment analysis of the pilot episode script, conducted using the Syuzhet package in R, quantifies the emotional intent derived from text written by screenwriters. Figure 1 shows that the three most frequent emotions in the script are trust (122 mentions), sadness (114 mentions), and fear (111 mentions). The prevalence of trust is directly linked to the design of the protagonist, lawyer Ana Tramel, who was conceived to foster a connection with viewers from the beginning.
Figure 1. The Emotional Design of the Pilot Episode Script.
Source: Own elaboration.
The script presents a remarkable balance of sentiment, although there is a slight predominance of negativity (218) over positivity (197), as shown in Figure 2. This tendency is consistent with the dramatic nature of the storylines, which include addiction, violence, and murder. Nevertheless, a significant balance between negative and positive feelings is evident despite the type of story told. Later, the perception of this proportional distribution by the audience will be examined.
Figure 2. Polarity of Sentiment in the Pilot Episode Script.
Source: Own elaboration.
This balance is evident throughout the narrative structure, as illustrated in Figure 3. The Y-axis represents sentiment value and the X-axis represents the line of dialogue. This allows for the determination of the emotional positivity or negativity of each sentence. Sentiment fluctuates constantly between positive and negative values throughout the nearly 1,400 lines of the script, preventing a monotonous tone and building dynamic emotional contrasts without affecting the overall balance of negative and positive feelings. These fluctuations influence audience attention and engagement levels (Diego et al., 2025).
Figure 3. Evolution of Sentiment Polarity in the Pilot Episode Script.
Source: Own elaboration.
After examining the emotional design of the script, the next step was to investigate the audience's emotional response in order to compare the two sets of results. The goal was to analyze the relationship between the emotional design of the content and emotional gratification. According to the initial hypothesis, emotional gratification fosters engagement. In other words, this study aimed to verify whether the production and writing teams of the series achieved the desired audience loyalty, which is a prerequisite for establishing a sustainable production model based on the quality and suitability of the content to the needs and sensitivities of diverse audiences.
4.2. The Audience's Emotional Response
Two complementary methods were used to analyze the emotional response: sentiment analysis of messages posted on Twitter during the live television broadcast with the official hashtag (#EstrenoAnaTramel), reflecting a more immediate and unconscious reaction, and a post-viewing questionnaire administered to 59 young people for a conscious and reflective evaluation.
The analysis of the messages yielded conclusive results. The three most frequently detected emotions in audience conversations (Figure 4) were fear (112), sadness (100), and trust (98). These are precisely the same three main emotions that emerged from the script, suggesting that the emotional design was effective in its transmission. However, the order of prevalence is altered: Fear, a typical emotion in the thriller genre, became the dominant emotion in the audience's reception, surpassing trust.
Figure 4. Emotions Derived From Audience Messages on Twitter (X).
Source: Own elaboration.
Figure 5 shows that the reaction on social media regarding sentiment polarity is much more pronounced than in the script. There is a strong predominance of negativity (166) over positivity (89), representing 65% of the total sentiments expressed. Therefore, the audience reaction is more polarized, a phenomenon typically associated with social media activity (Arce-García et al., 2022; Ruiz-Dodobara et al., 2024).
Figure 5. Sentiment Polarity in Audience Messages on Twitter (X).
Source: Own elaboration.
Figure 6 shows the evolution of this sentiment. The peaks of greatest negativity correspond to moments of maximum dramatic tension in the plot, which are particularly intense in the second half of the episode. These moments culminate in the apparent suicide of the character Alejandro Tramel, which marks the episode's climax and resolution. As can be seen, negativity dominates the evolution of sentiment in audience tweets throughout the broadcast; most viewers' messages exhibit negative values.
Figure 6. Evolution of Sentiment Polarity in Audience Tweets (X).
Source: Own elaboration.
To complement the results obtained through data-driven, computational sentiment analysis methods, a post-viewing questionnaire was administered to evaluate the audience's conscious perception. In this case, only those who participated in the private screening organized for this research were considered.
As previously mentioned, this study began with the conceptual distinction between emotion, which is an instinctive and immediate psychophysiological reaction, and feeling, which involves a subsequent cognitive evaluation of that emotion (Nünning, 2017). During this reflective phase, the emotional experience takes on a positive or negative value, known as the polarity of feeling (Mohammad & Turner, 2013).
Analysis of the conscious emotions evoked by the pilot episode (Figure 7) reveals that surprise was the predominant reaction, reported as intense or very intense by 46 of the 59 participants. Anticipation (34) and sadness (23) followed in relevance. In a thriller, it is not surprising that surprise stands out as the most recognized emotion by the audience. Regarding the polarity of sentiment, it was predominantly negative: 69.5% of viewers reported negative feelings, while 30.5% experienced positive ones. Nevertheless, the episode demonstrated remarkable engagement, with 76% of the panel stating that the plot captured their interest and expressing a desire to continue watching the series.
A detailed study of the reactions to the main characters offers the following insights:
- Ana Tramel, the protagonist, evokes an emotional response in which trust and sadness are equally dominant, as indicated by 38 out of 59 viewers. Significantly, despite the presence of sadness, the predominant polarity of the feeling elicited is positive (54% versus 46%). This apparent contradiction can be understood through the concept of empathic sadness. Rather than generating rejection, this mechanism fosters connection and emotional gratification through vicarious experience, which is key to engagement (Bartsch, 2012). This empathy is reinforced by the strength, courage, and resolve of the female character archetype (Labrador et al., 2023), which solidifies the emotion of trust due to the certainty with which the protagonist makes decisions.
- Alejandro, Concha, and Helena are other characters in the main plot who evoke a variety of emotions. Alejandro Tramel, the protagonist's brother and a pivotal figure in this episode, evokes sadness and surprise, with 38 mentions of each. These emotions are associated with an overwhelmingly negative sentiment (86%). Concha, the protagonist's business partner, generates an emotional profile similar to Ana's, based on trust and sadness. She receives an identical emotional rating (54% positive). Finally, Helena, Alejandro's partner, stands out for evoking sadness and surprise; however, the balance of feelings leans slightly toward the positive (52.5%).
- The case of Lieutenant Moncada, the antagonist, is unique. The emotions he conveys most strongly are trust, recognized by 31 people, and anticipation. His emotional profile is low and goes practically unnoticed by the selected sample. According to the questionnaire, this character hardly provokes any conscious emotions. This discretion, coupled with a positive sentiment polarity of 54%, aligns with the narrative strategy of concealing his role as the antagonist at the beginning of the series. The audience does not perceive him as a threat, instead attributing the trust inherent in his authority figure to him, thus fulfilling the screenwriters' objective.
Figure 7. Conscious Emotions and Feelings of the Audience Derived from the Questionnaire.
Source: Own elaboration.
In summary, combining the three methodologies enables a more thorough investigation. Script analysis reveals a balanced emotional design with trust in the protagonist as its main foundation. Unconscious social media reactions confirm the reception of key emotions (trust, sadness, and fear), albeit with a different hierarchy (fear becomes the primary emotion) and much more negative polarity linked to the plot's dramatic events. Finally, the conscious questionnaire corroborates the predominance of negative feelings (69.5%) and introduces emotions such as surprise and anticipation, which are fundamental to engagement with a thriller. This method also allows for the unraveling of complex phenomena, such as empathic sadness. This explains why a character like Ana Tramel can generate positive feelings despite evoking sadness. Thus, it validates the effectiveness of her emotional design.
5. DISCUSSION AND CONCLUSIONS
Based on the analysis of the results and in response to the stated objectives, this research yields several key conclusions about the relationship between the emotional design of screenplays and how audiences respond to audiovisual fiction.
First, the emotional design of the content is clearly effective, but it is transformed and prioritized in reception. The analysis shows that the three main emotions in the screenplay (trust, sadness, and fear) were also the most frequently detected in the audience's unconscious reactions on Twitter, validating the effectiveness of the narrative design. However, the order of importance was altered, prioritizing fear, which aligns with thriller genre conventions. However, when asked about their conscious emotions in the questionnaire, the audience reported surprise, another emotion characteristic of the genre. This finding underscores that while screenwriters can implement an emotional architecture, audiences decode and experience it through the filter of their expectations regarding the genre and thematic content.
Second, the emotional design of the characters, especially the protagonist, is fundamental to engagement and desire to watch the entire series after viewing the pilot. Although the series evokes predominantly negative feelings, high engagement (76%) is based on connection with the protagonist. Her complex emotional makeup (trust and empathetic sadness) generated a bond reinforced by positive sentiment (54%) that overcame the negativity of the plot. This suggests that dramatic twists can generate an immediate impact, as reflected in fluctuations of positive and negative sentiment in Twitter messages. However, it is identification with the characters that most effectively fosters audience loyalty.
Third, methodological triangulation is a useful tool for understanding the script's narrative design and the viewer's experience holistically. No single analysis would have offered such a comprehensive view. The script study reveals the creative intent and emotional structure, while the Twitter analysis captures the immediate impact and reveals how viewers react more negatively and polarized to dramatic peaks. Finally, the questionnaire offers a cognitive and reflective assessment of the audience's perceived emotions. In this case, the discrepancy in dominant emotions (trust in the script, fear on Twitter, and surprise in the questionnaire) is not contradictory, but rather evidence of a complex reception process. It shows that emotions are not perceived in isolation but complement each other, thereby enriching the viewing experience. Identifying these dissonances and synergies between designed content and perceived reactions can align creative intention with final audience reception in the various phases of production.
Fourth, analyzing the emotions of characters allows for the empirical validation of the effectiveness of narrative strategies, such as managing suspense or concealing a character's true nature. This study found that the antagonist generates trust, anticipation, and positive feelings while remaining almost imperceptible to the pilot audience. This confirms the screenwriters' success in masking the antagonist's role at the beginning of the series to avoid arousing suspicion. This is crucial for the development of serialized fiction because it gives creators an additional tool to ensure that plot twists have the desired effect.
Finally, aligning the script with the audience's emotional sensitivities provides showrunners with various opportunities from an executive production perspective. The combined application of these methodologies optimizes production policies, facilitating investment in quality projects that effectively connect with their target audience. This increases the likelihood of a series' success and contributes to the long-term sustainability of the production and business model by reducing uncertainty and fostering a more efficient, targeted content catalog.
Since this research methodology is applied exclusively to one episode from one season of a specific title, these conclusions are exploratory and have limitations. It would be pertinent to apply this analytical model to other genres and series with multiple seasons to validate the usefulness of this methodological combination.
In any case, these conclusions spark an interesting discussion for the industry and academic research. The relationship between unconscious emotional responses and conscious evaluations suggests the need to further study emotional gratification as a mechanism for audience engagement. In this regard, neuroscience can provide a complementary perspective, as demonstrated by the earlier cited studies.
Ultimately, this study contributes to the debate on using scientific methods and computational data analysis in audiovisual creation processes. Rather than limiting creativity, these tools reveal themselves to be valuable instruments that allow creators to better understand how to connect with audiences. This optimizes resources for more sustainable fiction production that resonates with audience sensibilities. Similarly, artificial intelligence presents an opportunity —though not without risks— to align the creative process with audience needs to improve content quality and likelihood of success.
6. REFERENCES
Amalia, A, Oktinas, W., Aulia, I., & Rahmat, R. F. (2018). Determination of quality television programmes based on sentiment analysis on Twitter. Journal of Physics: Conference Series, 978, 1-6. https://doi.org/10.1088/1742-6596/978/1/012117
Ángeles Gómez, D. I., & Quintana López, M. (2019). Análisis de sentimientos en videojuegos. ReCIBE, Revista electrónica de Computación, Informática, Biomédica y Electrónica, 8(2), 1-23. https://doi.org/10.32870/recibe.v8i2.137
Arce-García, S., Vila, F., & Fondevila-Gascón, J.-F. (2022). Analysis of the Twitter discourse in the 2019 electoral debates in Spain: a comparative algorithmic study. Communication & Society, 35(1), 45-61. https://doi.org/10.15581/003.35.1.45-61
Atarama-Rojas, T., & Feijoo, B. (2023). Social Audience and Emotional Bonding in Marvel’s Transmedia Phenomenon: An Exploration of Peruvian Digital Communities. International Journal of Communication, 17, 3054-3071. https://ijoc.org/index.php/ijoc/article/view/19603
Bartsch, A. (2012). Emotional Gratification in Entertainment Experience. Why Viewers of Movies and Television Series Find it Rewarding to Experience Emotions. Media Psychology, 15(3), 267-302. https://doi.org/10.1080/15213269.2012.693811
Bedi, G., & Tamrakar, A. (2019). Sentiment Analysis of a Movie Using Twitter and Text Analytics. International Journal Of Engineering Sciences & Research Technology, 8(4), 160-165. https://doi.org/10.5281/zenodo.2640752
Busselle, R., & Bilandzic, H. (2009). Measuring narrative engagement. Media Psychology, 12, 321-347. https://doi.org/10.1080/15213260903287259
Castro-Mariño, D. (2018). La audiencia social y el mundo narrativo de la ficción: análisis de los comentarios de los fans y community managers sobre la ficción televisiva y webseries españolas. Palabra Clave, 21(1), 86-110. https://doi.org/10.5294/pacla.2018.21.1.5
Cebral-Loureda, M., & Sued-Palmeiro, G. E. (2021). Análisis computacional de la conversación pública en lengua española. Cuadernos.info, 49, 1-25. https://doi.org/10.7764/cdi.49.27467
Ceron, A., & Splendore, S. (2018). From contents to comments: Social TV and perceived pluralism in political talk shows. New Media & Society, 20(2), 659-675. https://doi.org/10.1177/1461444816668187
Chicharro-Merayo, M., Gil-Gascón, F., & Arce-García, S. (2022). Narrativas de ficción para relatar una pandemia. Análisis y recepción de Diarios de la cuarentena (Televisión Española, 2020). Estudios sobre el Mensaje Periodístico, 28(2), 293-304. https://doi.org/10.5209/esmp.75946
Cuesta-Cambra, U., Niño-González, J.-I., & de-Marchis, G. (2021). Uses and gratifications of multiscreen news consumption among Spanish youth. Communication & Society, 34(2), 15-29. https://doi.org/10.15581/003.34.2.15-29
Datacamp (2025). Rdocumentation. https://www.rdocumentation.org/packages/syuzhet/
Diego, P. (2015). Estrategia multiplataforma de la televisión conectada. En: M. Medina, La audiencia en la era digital (pp. 145-168). Fragua.
Diego, P., Guerrero-Pérez, E., Martín-Guerra, E., & Romano, C. (2025). Decoding Audience Emotions. Applied Neuroscience in Fiction Series Production. In: M. Medina, & D. Kimber, Media Engagement. Connecting with Audiences in Media Markets (pp. 205-218). Routledge.
Doyle, G. (2016a). Digitization and Changing Windowing Strategies in the Television Industry: Negotiating New Windows on the World. Television & New Media, 7(7). https://doi.org/10.1177/1527476416641194
Doyle, G. (2016b). Television production, Funding Models and Exploitation of Content. Icono 14, 14(2), 75-96. https://doi.org/10.7195/ri14.v14i1.991
Doyle, G. (2018). Television production: configuring for sustainability in the digital era. Media, Culture & Society, 40(2), 285-295. https://doi.org/10.1177/0163443717717634
Ekman, P. (1972). Emotion in the human face: Guidelines for research and an integration of findings. Pergamon Press.
Ekman, P. (2007). Emotions revealed: recognizing faces and feelings to improve communication and emotional life. Henry Holt.
Etayo, C., Bayo-Moriones, A., & Sánchez-Tabernero, A. (2023). The Growth of the offer and the Perceptions of Television Content Quality. Journal of Media Business Studies, 20(1), 1-26. https://doi.org/10.1080/16522354.2021.1984159
Fernández-Gómez, E., & Martín-Quevedo, J. (2018). La estrategia de engagement de NetflixEspaña en Twitter. El Profesional de la Información, 27(6), 1292-1302. https://doi.org/10.3145/epi.2018.nov.12
Francés i Domènec, M. (2015). Hacia un nuevo consumo audiovisual: TV e Internet. In M. Francés i Domènec, G. Llorca Abad, & A. Peris Blanes, La televisión conectada en el entorno transmedia. Eunsa. https://ebooks.eunsa.es/library/publication/la-television-conectada-en-el-entorno-transmedia
GECA (2021). Link TV. https://www.geca.es/linktvnet
González-Bernal, M. I. (2016). Hacia la comprensión del engagement de las audiencias de televisión, modelo conceptual multidimensional desde la comunicación. Palabra Clave, 19(3), 769-809. https://doi.org/10.5294/pacla.2016.19.3.5
Guerrero-Pérez, E. (2018). La fuga de los millennials de la televisión lineal. Revista Latina de Comunicación Social, 73, 1231-1246. https://doi.org/10.4185/RLCS-2018-1304
Guerrero-Pérez, E., Diego, P., Romano de Pedro, C., & Martín-Guerra, E. (2025). Neurociencia aplicada a la producción de series de ficción: mejorar la calidad optimizando el engagement y los niveles de atención de la audiencia. Revista De Comunicación, 24(1), 265-280. https://doi.org/10.26441/RC24.1-2025-3663
Heath, R. (2007). How do we predict attention and engagement? En J. Plummer (Ed.), Nueva York: ARF Annual Convention (pp. 4-10). The Advertising Research Association, ARF.
Higueras-Ruiz, M. J., Gómez-Pérez, F. J., & Alberich-Pascual, J. (2021). The showrunner’s skills and responsibilities in the creation and production process of fiction series in the contemporary North American television industry. Communication & Society, 34(4), 185-200. https://doi.org/10.15581/003.34.4.185-200
Holt, J., & Sanson, K. (2014). Connected Viewing. Selling, Streaming & Sharing Media in the Digital Era. Routledge. https://dx.doi.org/10.5209/esmp.75946
Jenkins, H. (2006). Convergence Culture: Where Old and New Media Collide. New York University Press.
Jockers, M. L. (2015). Syuzhet: Extract Sentiment and Plot Arcs from Text. https://cran.r-project.org/package=syuzhet
Katz, E., Blumler, J. G., & Gurevitch, M. (1973). Uses and gratifications research. Public Opinion Quarterly, 37(4), 509. https://doi.org/10.1086/268109
Kimber-Camussetti, D., & Guerrero-Pérez, E. (2022). Production values as program quality signals in Spanish linear TV: A comparison of two periods. Communication & Society, 35(2), 107-120. https://doi.org/10.15581/003.35.2.107-120
Kiritchenko, S., Zhu X., & Mohammad, S. M. (2014). Sentiment Analysis of Short Informal Texts. Journal of Artificial Intelligence Research, 50, 723-762. https://doi.org/10.1613/jair.4272
Labrador, M. J., Arrojo, M. J., Martín, M. E., Romano, C., Nicolás, M. T., & Vidal, C. (2023). Mujeres. Análisis emocional en la ficción seriada multiplataforma. Tirant lo Blanch.
López Villanueva, J. (2011). La reconfiguración de la cadena de valor. En: J. M. Álvarez Monzoncillo, La televisión etiquetada: nuevas audiencias, nuevos negocios. Ariel. https://www.academia.edu/5363754/LA_TELEVISION_ETIQUETADA
Malik, H., Shakshuki, E. M., & Yasar, A. U. H. (2022). Approximating viewership of streaming TV programs using social media sentiment analysis. Procedia Computer Science, 198, 94-101. https://doi.org/10.1016/j.procs.2021.12.215
Mohammad, S. M., & Turney, P. D. (2013). Crowdsourcing a Word-Emotion Association Lexicon. Computational Intelligence, 29(3), 436-465. https://doi.org/10.1111/j.1467-8640.2012.00460.x
Nünning, V. (2017). The Affective Value of Fiction. Presenting and evoking emotions. In I. Jandl, S. Knaller, S. Schönfellner, & G. Tockner, Writing Emotions, Transcript Verlag (pp. 29-54). https://doi.org/10.2307/j.ctv1wxt3t.5
Plutchik, R. (1962). The Emotions. Random House.
Plutchik, R. (1980). A general psychoevolutionary theory of emotion. Emotion: Theory, Research, and Experience, 1(3), 3-33.
Plutchik, R. (1994). The Psychology and Biology ofEmotion. Harper Collins.
Quintas-Frofe, N., & González Neira, A. (2021). Los estudios de la audiencia. De la tradición a la innovación. Editorial Gedisa.
Rodríguez Vázquez, A. I., García-Orosa, B., Portilla, I., & Direito-Rebollal, S. (2019). El panorama cambiante de los sistemas de medición de audiencias. Espejo De Monografías De Comunicación Social, 3, 201-223. https://doi.org/10.52495/c9.emcs.3.p73
Ruiz-Dodobara, F., Uribe-Bravo, K.-A., & Chaparro, H. (2024). Social Media, Affective Polarization, and Collective Action in Peru. Communication & Society, 37(4), 39-55. https://doi.org/10.15581/003.37.4.39-55
Saló, G., Guerrero, E., Formoso, M. J., & Maestro, S. (2022). La élite de las OTT: la medición de audiencias bajo demanda del público joven. In: P. Sidorenko Bautista, J. M. Herranz de la Casa, R. Terol Bolinches, & N. Alonso López, Narrativas emergentes para la comunicación digital (pp. 295-324). Dykinson. https://www.dykinson.com/libros/narrativas-emergentes-para-la-comunicacion-digital-2-vols/9788411223751/
Tapia Frade, A., Martín Guerra, E., & Romano de Pedro, C. (2025). Atención y emoción en spots televisivos. Un análisis desde la neurociencia. Vivat Academia, 158, 1-20. https://doi.org/10.15178/va.2025.158.e1606
Vizjak, A., & Ringlstetter, M. J. (2003). Media management: Leveraging content for profitable growth. Springer. https://link.springer.com/book/10.1007/978-3-540-24786-9?page=1#toc
Vonderau, P. (2015). The Politics of Content Aggregation. Television & New Media, 16(8), 717-733. https://doi.org/10.1177/1527476414554402
Zillmann, D. (1988). Mood management through communication choices. American Behavioral Scientist, 31, 327-340. https://doi.org/10.1177/000276488031003005
Zillmann, D. (1996). The psychology of suspense in dramatic exposition. In: P. Vorderer, H. J. Wulff, & M. Friedrichsen (eds.), Suspense: Conceptualizations, theoretical analyses, and empirical explorations (199-231). Erlbaum. https://www.routledge.com/Suspense-Conceptualizations-Theoretical-Analyses-and-Empirical-Explorations/Vorderer-Wulff-Friedrichsen/p/book/9780805819663
CONTRIBUTIONS OF AUTHORS, FINANCING AND ACKNOWLEDGMENTS
Contributions of Authors:
Conceptualization: Enrique Guerrero-Pérez and Patricia Diego. Software: Enrique Guerrero-Pérez. Validation: Enrique Guerrero-Pérez and Patricia Diego. Formal Analysis: Enrique Guerrero-Pérez and Patricia Diego. Curation of data: Guerrero-Pérez, Enrique and Diego, Patricia. Drafting-Preparation of the draft original: Guerrero-Pérez, Enrique and Diego, Patricia. Drafting -Revision and Edition: Guerrero-Pérez, Enrique and Diego, Patricia. Visualization: Guerrero-Pérez, Enrique and Diego, Patricia. Supervision: Guerrero-Pérez, Enrique and Diego, Patricia. Project Management: Guerrero-Pérez, Enrique and Diego, Patricia. All The authors have read and accepted the published version of the manuscript: Guerrero-Pérez, Enrique and Diego, Patricia.
Funding: Research project “Quality as an essential factor for a sustainable business model in streaming platforms”, funded by the Ministry of Science, Innovation and Universities of the Government of Spain. Start date: 09/01/2024. End date: 08/31/2027. Reference PID2023-150258NB-I00.
AUTHOR(S):
Enrique Guerrero-Pérez
Universidad de Navarra.
Professor in the Faculty of Communication with a Doctorate in Communication and an Extraordinary Doctoral Award. His research focuses on the production of entertainment content, content management strategies, and the impact of innovation on quality and sustainability policies in the audiovisual streaming sector. He is the author of El entretenimiento en la televisión española (Entertainment on Spanish Television) (Deusto) and Guion y producción de programas de entretenimiento (Scriptwriting and Production of Entertainment Programs) (Eunsa). He is also the principal investigator of a research project funded by the Spanish Ministry of Science, Innovation, and Universities. His professional experience includes research stays at the University of California, Los Angeles (UCLA), Bournemouth University (UK), and the University of Texas at Austin (USA).
Orcid : https://orcid.org/0000-0001-7693-8669
Google Scholar: https://scholar.google.com/citations?user=wwf7ZOMAAAAJ&hl=es
Patricia Diego
Universidad de Navarra.
She is a professor at the Faculty of Communication at the University of Navarra (UNAV). She has a Doctorate in Communication and was awarded the Extraordinary Doctoral Prize. Her research focuses on the production of fictional series and content management policies for streaming fiction. She is the author of the book La ficción en la pequeña pantalla (Fiction on the Small Screen) (Eunsa) and a member of the Spanish Television Academy. She has been a lead researcher and has participated in several publicly funded research projects. Her international experience includes visiting research positions at prestigious institutions such as the British Film Institute (UK), the University of Westminster (UK), and University College Cork (Ireland).
Orcid : https://orcid.org/0000-0002-7616-2474
Google Scholar: https://scholar.google.com/citations?user=MJdlScEAAAAJ&hl=es