Executive summary

Analyzing the online activity of candidates between April 24 and May 11, 2025, we discovered artificial amplification tactics that go beyond legitimate promotion. Our analysis identified 286 accounts with suspicious behavior for George Simion and 135 for Nicușor Dan, both disseminating identical and repetitive messages. Although the first electoral round on May 5 partially explains the increases in engagement, the presence of commenters from countries with no ties to Romania and the use of rare languages such as Amharic or Khmer indicates a deliberate strategy of inauthentic engagement. With just three days before the second round of the presidential elections, this report provides the public with relevant information about the reality behind the spectacular increases in online popularity and the invisible factors that can influence public perception.

Electoral context and approach

Funky Citizens has conducted a detailed analysis of the online activity of presidential candidates on Meta platforms (Facebook and Instagram) and TikTok. The analysis continues a first edition that explored the online activity of candidates in the first part of the electoral campaign (April 1-23, 2025). In this report, we continue monitoring social media presence during the period April 24 - May 11, with an extension of the analysis to accounts that commented on the official profiles of candidates during the period April 1 - May 11. The report aims to decipher the digital strategies used by candidates in online promotion, content typology, and to provide an assessment of the impact of campaign messages transmitted by candidates through social networking platforms.

By examining their official profiles, we sought to identify both distinctive elements and common trends in how presidential candidates build their image and communicate with the electorate in the online environment. Moreover, the extension of the analysis to comments left on candidate accounts provides a perspective on suspicions of inauthentic behavior in promoting candidates through accounts from foreign countries or profiles without content and activity on platforms.

Why is it important to look at candidate accounts?

Recently, discussions about the inauthentic promotion of candidates have intensified. Several investigations conducted during the electoral campaign show that there is inauthentic activity on social networking platforms when it comes to promoting candidates.

However, the key moment is the campaign conducted during the first round of the presidential elections by Călin Georgescu, the surprise candidate of the November 2024 elections, in which massive increases were identified on social networking platforms, especially TikTok. We recall that this campaign propelled Călin Georgescu, a marginal figure previously known for his anti-Western opinions, towards an unexpected victory in the first round through what appeared to be a carefully orchestrated operation in social media. Intelligence services officially accused the existence of a campaign worth 1 million euros, involving 25,000 coordinated TikTok accounts, none of which were declared in the official campaign financing reports. According to investigations, the operation presented clear characteristics of a state actor, with technical infrastructure linked to Russia supporting the distribution of content in 19 countries.

Funky Citizens published in December 2024 a final report regarding the disinformation campaigns from the first round of the presidential elections, which can be consulted in detail here - Undermining democracy: Weaponizing Social Media in Romania's 2024 elections.

In the campaign for the 2025 presidential elections, another key moment that sends worrying signals regarding the propulsion of candidates on social networks is the inauthentic growth of the official pages of Nicușor Dan, an independent presidential candidate. The candidate publicly denounced what he described as a coordinated attack against his social media accounts. During a press conference, he declared that they were "invaded by hundreds of thousands of fake accounts" from countries such as Vietnam, India, and Pakistan, forcing him to suspend the online campaign, for which he was spending approximately 15,000 euros daily. Along with Nicușor Dan, former presidential candidate Victor Ponta denounced a similar inauthentic growth attack a week before the first round.

Given the context of the 2025 presidential elections and the unprecedented events that led to the cancellation of the 2024 elections, we consider it necessary to pay increased attention to how candidates in this round manage to reach the electorate and to identify, where appropriate, anomalies in unexpected increases that propel candidates overnight. Funky Citizens continues to assume close monitoring of false narratives circulating during this period on social networks, along with careful monitoring of candidates' electoral campaigns to inform the public about the activity of candidates they are about to vote for in less than a few days.

How does the information landscape look on platforms near the presidential elections?

If you want to find out more about the evolution of disinformation near the presidential elections, you can consult the weekly newsletters available here: Romania's information landscape in the context of the 2025 presidential elections.

The information landscape on social media platforms around the Romanian presidential elections is dominated by a complex network of disinformation. According to our analysis of reports, we identified several predominant trends and narratives:

The main narratives include:

• Călin Georgescu and George Simion as "incumbent presidents": Posts presenting these politicians as already invested or as viable candidates for the current electoral process, supported by misleading and emotionally charged content.
• Accusations of electoral fraud: Unconfirmed claims about vote manipulation, falsifying ballots, and "ghost" voters, accompanied by warnings about planned power outages ("EIC: Blackout") and misinformation regarding voting stamps ("EIC: Stamp").
• Cancellation/Delay of elections: Narratives predicting or promoting the cancellation of the electoral process if the "wrong" candidate wins, based on theories about alleged foreign interventions or emergency declarations.
• Government crimes and coups d'état: Accusations that the Romanian government is committing crimes against its own citizens and organizing coups by canceling the electoral process, using dramatic and alarmist language.
• Calls for action against the constitutional order: Messages encouraging national mobilization and violent resistance against authorities, presented as necessary for "restoring constitutional order".
• Premature claims regarding polls: Presentation of unofficial and fabricated figures about voting results, exaggerating support for certain candidates.

Regarding conspiracy theories, the following narratives predominate:

• Parallel state: Narratives about the existence of a "parallel state" plotting against populist candidates, involving intelligence services, the judicial system, and political leaders.
• EU subjugating Romania: Conspiracy theories accusing "globalist oligarchs", often linked to the EU or France, of manipulating Romania toward war or ruin.
• Jewish plot against Romania: Anti-Semitic theories attributing genocidal plans to figures like George or Alex Soros.

Defamation campaigns:

• Intensification of narratives attempting to denigrate Nicușor Dan through artificial intelligence-generated content, speeches, memes, and deepfakes.
• Similar campaigns also target other candidates, especially Crin Antonescu, presenting them as part of an "anti-Georgescu political bloc" or accomplices in a wider anti-national scheme.

During the analyzed period, we identified inauthentic coordinated activity aimed at promoting and amplifying the messages of populist candidates and parties. This activity takes place through networks of fake or coordinated managed accounts, with the purpose of manipulating public perception.

Methodology

To create a broader yet detailed perspective of the increases identified among candidates, we focused on the main indicators of account growth on platforms: number of followers, likes, shares, and engagement statistics for posts and official profiles of candidates. The entire analysis is specifically centered on the official accounts of presidential candidates, precisely to determine if there are amplifications and popularity peaks on social networking platforms during the period April 1-23, 2025.

However, a few mentions must be made to ensure the accuracy of analyses based on collected data:

• The analysis focuses on all posts made by candidates during the monitored period on Meta and TikTok platforms (the most used in Romania);
• Interaction indicators (views, likes, comments, and shares) are analyzed within the access granted by platforms - there are cases which are detailed below, where data about views or likes are not available;
• Day-to-day fluctuations of engagement metrics (likes, comments, shares, views);
• Statistical anomalies in engagement reports;
• Linguistic analysis of suspicious patterns in comments;
• Temporal distribution of engagement (frequency of comments per minute);
• Detection of language of accounts among followers;
• Cross-platform comparisons for identifying synchronized patterns;
• Comparative analysis with statistical norms of organic behavior on each platform.

We also analyzed in detail user comments on posts disseminated through their official accounts. Thus, we collected a total of 69,529 comments related to the posts of the two candidates. Specifically, for Nicușor Dan we examined 92 posts that generated 30,997 comments from 13,076 unique users, while for George Simion we analyzed only 25 posts that attracted a significantly higher number of comments - 38,532, coming from 24,276 unique users. We remind that the analysis was conducted based on comments disseminated during the period April 1 - May 11, 2025, on the posts of the official accounts of candidates on the TikTok platform.

In the analysis, we evaluated potential inauthentic behavior present on the social accounts of candidates Nicușor Dan and George Simion. Our methodology was based on five main indicators:

• Comment frequency - the number of comments left by each user;
• Duplication rate - the degree of repetitiveness in comment content;
• Activity distribution - the number of identical or very similar texts that a user has disseminated in comments;
• Time intervals in which comments were posted;
• Textual data derived from comments.

These indicators were identified following a review of specialized literature and the identification of a similar study. Thus, we proceeded to the descriptive analysis of indicators related to comments in the corpus and identified and aggregated relevant indicators for creating a score that would indicate the probability of inauthentic behavior of accounts, based on information strictly derived from those comments. Based on these parameters, we developed the indicator called "botscore", which quantifies the degree of inauthentic behavior and applied it to each comment, thus identifying specific characteristics and trends for each comment corpus, related to candidate posts.

Analysis tools and techniques

For this digital investigation, we used:

• Social media monitoring tools: Specialized systems for collecting data on engagement, views, and shares in real-time.
• Analysis tools: Tools and programming languages for modeling, data visualization, and statistical analysis.
• Anomaly detection algorithms: Statistical methods for identifying significant deviations from normal patterns analyzed.
• Linguistic analysis: Techniques and algorithms for natural language analysis to identify relevant patterns associated with inauthentic behaviors.
• Temporal analysis: Examining the temporal distribution of comments to identify behaviors suggesting automation.
• Network analysis: Techniques for identifying potential connections between accounts involved in suspicious activities.

All these methods were applied with careful attention to data confidentiality, using only public information available on the respective platforms.

Methodology limitations

While our methodology is robust, we acknowledge several important limitations:

• Limited access to data: Platforms restrict access to complete data, especially for private accounts or internal metrics.
• Algorithmic changes: Frequent changes in platform algorithms can influence visibility and engagement in a way that does not necessarily reflect manipulation.
• Legitimate marketing interventions: Some peaks may result from legitimate digital marketing campaigns, although the specific patterns suggested in the report are atypical even for paid promotion.
• External factors: Current events can influence engagement without necessarily implying manipulation.

The model used for creating the "botscore" also presents limitations, as we identified 687 unique inauthentic accounts deleted at intervals other than those identified as statistically relevant. The respective accounts repeated only 3 texts in the corpus of posts of candidate "Nicușor Dan": "The only correct candidate" (111 entries), "One correct candidate" (313 entries), and "The only correct candidate" (290 entries), totaling 714 messages. Some accounts exhibiting inauthentic behavior were not detected through the current methodology due to the restricted dataset, which did not allow us to include all posts where these accounts interacted through comments, thus significantly reducing detection chances. To counteract this limitation, we also included in our analysis an analysis of user distribution according to the country of account registration and the language used in the comment.

These limitations were taken into consideration in the interpretation of data and formulation of conclusions.

Cross-platform patterns: "the smoking gun"

The most convincing evidence of possible digital manipulation becomes visible when we analyze engagement patterns between platforms. The patterns identified in the second part of the campaign confirm suspicions raised in the previous analysis:

Synchronized growth peaks in the period May 3-5, 2025 are simultaneously visible for both main candidates on all monitored platforms. This synchronization, especially between TikTok and Facebook, suggests strategically coordinated multi-platform amplification campaigns. It is noteworthy that George Simion records an astronomical peak of 10 million likes on TikTok on exactly the same day (May 5) when Nicușor Dan reaches his maximum of 6 million.

Significant differences appear, however, in the linguistic and geographical distribution of interactions. While George Simion's comments reflect a suspicious linguistic diversity, including rare languages such as Amharic, Central Khmer, and other languages, far beyond areas with Romanian diaspora, the analysis of Nicușor Dan's audience reveals a different, but equally problematic behavior - a restricted, but extremely active group of accounts with abnormally high "botscore" scores.

The invisible army: Anatomy of digital manipulation

An indicator of possible manipulation is the presence of accounts with inauthentic behavior. The detailed analysis of comments reveals two distinct approaches:

George Simion: A strategy of volume and geographical diversity. The identification of 286 accounts with abnormal activity that together generated 492 comments represents a discrete but efficient "digital army", with an average of 1.72 comments per account. These accounts stand out through the consistency of messages and their suspicious distribution in countries with no obvious connections to Romanian politics.

Nicușor Dan: A tactic of concentrated intensity. The analysis identified only 135 accounts with atypical behavior, but these generated an impressive 644 comments - an average of 4.7 comments per account. These "amplification nodes" are fewer, but much more active and generate extreme "botscore" scores, over 0.6, indicating a quite high probability of automated behavior.

Strategic narratives: The language of digital manipulation

The textual analysis of the most frequent comments reveals distinct communication strategies for each candidate:

For Nicușor Dan, repetitive comments focus on legitimacy and competence: "the only correct candidate", "one correct candidate", and "the only correct candidate" represent the overwhelming majority of repetitive messages. In the case of George Simion, recurring comments leverage religious and familiar connotations: "God bless you", "Much health to you and your family", creating an emotional and personalized connection with the audience.

These patterns are not just linguistic preferences - they represent deliberate strategies of positioning and creating a dominant narrative around each candidate.

What exactly is inauthentic behavior on platforms?

Inauthentic behavior on social platforms refers to coordinated activities that simulate genuine human interactions but are actually artificially generated or strategically manipulated to amplify the visibility of certain content or personalities.

What are the signs?

• Use of fake accounts or bots: Automated or semi-automated accounts that simulate human interactions to generate an artificial volume of engagement (likes, comments, shares).
• Coordinated engagement: Groups of accounts that act synchronously, generating interactions within a short time interval, thus creating the impression of spontaneous popularity.
• Repetitive or generic comments: Comments with similar or identical content posted by different accounts, often with vague formulations that can be applied to any type of content.
• Sudden and unjustified increases: Abnormal jumps in the number of followers or in engagement that cannot be explained by possible current events or by legitimate marketing strategies.
• Atypical geographical distribution: Abnormal concentration of followers or interactions from unexpected geographical areas for the respective context and content.

Analysis regarding inauthentic behavior on TikTok

On both graphs, the vertical axis indicates a score between 0 and 0.63, representing the probability that a comment is generated by an account exhibiting inauthentic behavior. The higher the score, the greater the suspicion of inauthentic activity.

The boxplots below illustrate the distribution of scores for comments related to the posts of the two candidates — Nicușor Dan and George Simion — revealing different patterns of digital behavior, with implications for the authenticity of interactions. Despite the fact that the median score is identical for both candidates (approximately 0.20), differences appear in how scores are distributed around this median and in the number of outliers identified.

For Nicușor Dan, the distribution is more "compact" around suspicious values, with a first and third quartile (Q1 and Q3) higher than those of George Simion (Q1 ≈ 0.13 and Q3 ≈ 0.27). This indicates a greater concentration of scores in moderate intervals, which may suggest a constant volume of comments with potential inauthentic character. In addition, although the total volume of comments is smaller than in the case of George Simion's posts, Nicușor Dan's posts record a higher number of outliers — 644 compared to 492 — a sign that there are more comments with bot scores considerably higher than the average distribution.

Fig.1. Boxplot of comments on Nicușor Dan's posts. 

On the other hand, the distribution of scores for comments from George Simion's posts is more widespread, with a larger difference between Q1 (≈ 0.10) and Q3 (≈ 0.25), which indicates a greater variety of user behaviors. Although this candidate has a larger volume of comments and unique users on a smaller number of posts, he records fewer outliers. This may signal a more diffuse and possibly more subtle distribution of automated activity, or more efficiently integrated into online conversations.

Fig.2. Boxplot of comments on George Simion's posts. 

Interpreting these observations in context, we note that Nicușor Dan has almost four times more posts, but attracts fewer comments and unique users. Despite this fact, he has more outliers, which may indicate either a more aggressive attempt at digital manipulation, or a weaker visibility that amplifies the impact of inauthentic comments. In contrast, George Simion benefits from a greater mobilization of users (almost twice as many unique commenters), but with a more uniform and, apparently, better dissimulated suspicious presence.

In conclusion, the data suggests two distinct patterns of inauthentic behavior: on Nicușor Dan's posts, a denser and more concentrated suspicious presence, with numerous outliers indicating peaks of automated activity; and on George Simion's posts, a more widely distributed suspicious activity, but with a relatively smaller impact in terms of extreme scores.

What do people comment on the posts of the two candidates?

We analyzed the top 20 texts that are most widespread in the comments section from both candidates. In the representation below, we included the content of the most widespread comments, along with the frequency reported in the analyzed dataset.

Comments for Nicușor Dan predominantly focus on the idea of competence and correctness ("correct candidate", "the only correct candidate", "the only correct candidate"), the latter being flagged as generated by bots. His supporters frequently use direct phrases of electoral support ("I vote for Nicușor Dan", "we vote for Nicușor Dan").

In contrast, comments for George Simion contain more expressions with religious and personal wishes ("God help", "God help you", "much health", "health", "health"), along with manifestations of encouragement and respect ("bravo", "respect", "congratulations").

Comment analysis

George Simion

• 286 accounts signaling abnormal behavior. The analysis of activity on George Simion's TikTok page highlighted a distinct group of 286 accounts that exhibit atypical engagement behavior. These accounts together generated 492 comments distributed across the 25 posts examined.

For data accuracy, three mentions are relevant for contextualizing the analysis: 1) A video's audience is calculated based on a subset of the public who commented on the video; 2) Languages refer to the language used in the comment; 3) Countries refer to the location of identified commenters.

How is the audience distributed by countries?

Audience distribution with Romania included

Romania occupies a dominant position in George Simion's audience structure, representing approximately 80% of the total people who interact through comments with his content. The next positions are occupied by Italy, United Kingdom, Germany, Spain, France, and Austria, showing the presence of the candidate's Romanian diaspora audience.

Audience distribution excluding Romania

To provide a more nuanced perspective on the main countries where the candidate's messages reach during the electoral campaign, we analyzed the distribution of countries excluding Romania to observe the high presence of the candidate in foreign countries. According to statistics from 2021-2022, the largest Romanian communities abroad are in Spain (1,087,923), Italy (1,083,771), Germany (844,535), USA (448,307) and the United Kingdom (342,000), Israel (220,000), Canada (215,885), Austria (140,454), France (132,291) and Belgium (106,293).

This geographical distribution largely corresponds to the model of George Simion's audience identified in the previous analysis. The countries with the largest Romanian communities - Italy, United Kingdom, Germany, Spain, France, Austria, and Belgium - are exactly those mentioned as having a significant presence in interactions with his content. This correlation suggests that Simion manages to attract the attention of Romanians from the main diaspora communities.

Distribution of languages spoken by George Simion's audience with Romanian language included

The Romanian language dominates the audience, representing approximately 85% of the total. The rest of the linguistic distribution includes mainly English, French, and Italian, followed by Spanish and Chinese with lower percentages. The presence of languages such as Indonesian, Turkish, Amharic, and many others (including rare languages such as Hausa, Central Khmer, and others) raises questions about the authenticity of some interactions. Although the Romanian diaspora is concentrated in countries such as Italy, Spain, and Germany, the unusual diversity of identified languages, extending far beyond known areas with Romanian communities, may suggest possible artificial or inauthentic interactions - an additional indicator that corroborates the high score of inauthentic behavior (botscore) identified for George Simion's account.

Distribution of languages spoken by George Simion's audience without Romanian language

The English language dominates with approximately 4.3% of the total, followed closely by French (approximately 4%). Italian occupies the third place with about 2.8%, and Spanish represents approximately 1.5%. Surprisingly, Chinese appears with approximately 1.2%, and Portuguese with approximately 1%. The other identified languages - Indonesian, Turkish, Amharic, Dutch/Flemish, Japanese, Malay, Tagalog, Vietnamese, Azerbaijani, and others - have percentages below 1%.

Nicușor Dan

• 135 accounts signaling abnormal behavior The analysis of activity on Nicușor Dan's TikTok page highlighted a distinct group of 135 accounts that exhibit atypical engagement behavior. These accounts together generated 644 comments distributed across the 95 posts examined.

Distribution of languages spoken by Nicușor Dan's audience with Romanian language included

The graph presents a significantly different linguistic distribution in the case of Nicușor Dan's audience compared to George Simion. Although Romanian (79.84%) remains predominant, its share is reduced, indicating a proportionally larger international audience. Surprisingly, Indonesian (3.39%) occupies the second place, followed closely by Italian (3.31%), English (3.16%), and French (3.14%), while German represents 2.36% and Spanish only 1.28%. The presence of major European languages partially corresponds to the distribution of the Romanian diaspora, but the positioning of Indonesian in second place and the presence of languages such as Chinese (0.91%), Vietnamese (0.34%), and Corsican (0.26%) raise questions about the authenticity of some interactions. Compared to Simion, Nicușor Dan's distribution presents fewer rare languages, and traditional diaspora languages (Italian, English, French, German) have more balanced shares.

Distribution of languages spoken by Nicușor Dan's audience without Romanian language

Analyzing Nicușor Dan's non-Romanian audience, we observe a relatively balanced distribution among the main foreign languages. Indonesian surprisingly leads with 3.39%, surpassing traditional diaspora languages such as Italian (3.31%), English (3.16%), and French (3.14%). German represents 2.36%, and Spanish only 1.28%, although Romania has numerous communities in both countries. Dutch/Flemish (0.97%), Chinese (0.91%), and Portuguese (0.54%) complete the picture of the main foreign languages. However, compared to George Simion's profile, Nicușor Dan's distribution presents fewer exotic languages and a greater concentration on the languages of countries with Romanian diaspora.

Audience distribution with Romania included

Romania dominates with 81.23% of the audience, slightly higher than in the case of George Simion (approximately 80%), which indicates a predominantly internal base of supporters. Italy occupies the second place with 5.5%, followed by the United Kingdom (3.53%), Germany (3.1%), and Spain (2.18%) - countries that coincide with the main destinations of the Romanian diaspora. China surprisingly appears in sixth place (1.01%), ahead of European countries with traditional Romanian communities such as Belgium (0.67%), France (0.64%), or Austria (0.43%). The Republic of Moldova has a modest presence (0.54%), and the USA represents only 0.26% of the audience, although it hosts the fourth largest Romanian community in the diaspora.

Audience distribution excluding Romania

Excluding Romania from the analysis, we can more clearly observe the international distribution of Nicușor Dan's audience. Italy categorically dominates with 5.5% of his total audience. The United Kingdom occupies the second place with 3.53%, followed by Germany with 3.1% and Spain with 2.18% - all these being countries with significant Romanian communities. The presence of China in fifth place (1.01%) represents a notable anomaly, being the only non-European country with a significant share and no obvious demographic connection to Romania. Belgium (0.67%), France (0.64%), and the Republic of Moldova (0.54%) complete the group of countries with a moderate presence. Austria, although hosting an important Romanian community, has a lower representation (0.43%).

Social Media presence

Facebook

Followers

Analysis of the evolution of the number of followers on Facebook: George Simion vs. Nicușor Dan

George Simion: Undisputed leader with a solid base in continuous growth George Simion demonstrates a dominant presence on Facebook, with an impressive base of 1,500,000 followers at the beginning of the analyzed period (April 23, 2025). Remarkably, this significant figure remains stable until May 2, only to then register a notable increase, reaching 1,600,000 followers by May 12. This evolution represents an advance of 100,000 new followers (approximately 6.7%) in the analyzed period, consolidating his dominant position in the political space on Facebook.

Nicușor Dan: Constant growth from a modest base Nicușor Dan starts from a considerably more modest position on Facebook, with 397,000 followers on April 23. His evolution shows, however, a constant positive dynamic, reaching 405,000 on May 2 and continuing growth to 485,000 on May 12. This progress represents a total advance of 88,000 followers (approximately 22.2%) in the analyzed period. Although his digital community remains more than three times smaller than that of George Simion, the proportionally more accelerated pace of growth suggests an efficiency increase of his communication strategy on this platform.

Likes

George Simion: Evident supremacy and spectacular peak on Facebook George Simion demonstrates a dominant presence on Facebook, with a considerably higher number of likes than his political rival for most of the analyzed period. The graph highlights constant fluctuations with a generally ascending tendency, culminating with an impressive peak on May 4-5 when he reaches approximately 410,000 likes. The end of the period shows a strong comeback, with values around 310,000 likes on May 11. These figures suggest that Facebook represents a strategic platform for George Simion, where he enjoys an engaged audience receptive to his messages.

Nicușor Dan: Modest evolution with punctual moments of impact Nicușor Dan presents an initially modest performance on Facebook, with values that generally fluctuate below 50,000 likes until the beginning of May. The situation changes significantly starting with May 3, when a sudden increase is observed that culminates with a peak of approximately 260,000 likes on May 5 - an impressive value, although inferior to the maximum recorded by Simion. After a temporary decrease, the end of the period marks a comeback, with an ascending tendency toward values of 200,000 likes. This dynamic suggests a potential strategic recalibration of his presence on this platform.

Comments

Comparative analysis of comments on Facebook: George Simion vs. Nicușor Dan George Simion demonstrates a superior performance regarding comments on Facebook, with significant fluctuations but with a general ascending tendency throughout the analyzed period. The graph highlights notable peaks on April 28 and May 2 (approximately 27,000 comments), followed by a spectacular increase that culminates with an impressive maximum of approximately 88,000 comments on May 4. A second important peak is observed on May 6 (approximately 70,000 comments), followed by a stabilization at values between 25,000-44,000 comments toward the end of the period. Nicușor Dan presents a considerably reduced performance in terms of comments, with values that fluctuate below 16,000 in the first half of the analyzed period. A notable improvement is observed on May 5, when he reaches approximately 26,000 comments, followed by a return to more modest values. The end of the period marks, however, a constantly ascending tendency, with values that approach 27,000 comments on May 11. Although he remains at a level inferior to his political rival, this evolution suggests a gradual increase in the ability to generate conversations around his content.

Analysis of comments on Facebook: George Simion vs. Nicușor Dan

George Simion presents a dynamic evolution of shares on Facebook, marked by considerable variations throughout the analyzed period. After a first week with moderate values (below 7,000), a first important peak is observed on May 2 (approximately 26,000 shares), followed by two spectacular maximums: one on May 4 and another on May 6, both reaching approximately 40,000 shares. After a sudden decrease at the beginning of the second week of May, the end of the period marks a strong comeback, with an ascending tendency that reaches approximately 37,000 shares on May 11.

Nicușor Dan begins the period with modest values of shares (below 7,000), recording minor fluctuations in the first part of the analyzed interval. A notable improvement is observed starting with May 4, culminating with a peak of approximately 22,000 shares on May 5. After a temporary decrease, the last part of the period highlights the most spectacular evolution, with a constant and accelerated increase that reaches approximately 34,800 shares on May 10, followed by a slight decrease to approximately 30,000 on May 11.

Analysis of shares on Facebook: George Simion vs. Nicușor Dan

Instagram

Comparative analysis of followers on Instagram: Nicușor Dan vs. George Simion

Nicușor Dan: Clear domination on Instagram Nicușor Dan stands out through an impressive presence on Instagram, with a substantial and continuously growing digital community. The data shows a constant evolution of his follower base: 585,000 on April 23, 667,000 on May 2 (an increase of 82,000 followers), and 712,000 on May 12 (an additional advance of 45,000). This ascending trajectory, with a total growth of 127,000 followers (approximately 22%) in the analyzed period, confirms the constant attractiveness of his content for the audience on this platform. Instagram proves to be the force communication channel for Nicușor Dan, where he holds a clear competitive advantage.

George Simion: Modest presence, in slight growth George Simion records a significantly reduced performance on Instagram compared to his political rival. His follower base, although growing, remains at a much inferior level: 52,900 on April 23, 53,900 on May 2 (a minimal increase of only 1,000 followers), and 62,800 on May 12 (a more consistent advance of 8,900). Although an acceleration of the growth rate is observed in the last analyzed period, the difference compared to Nicușor Dan remains overwhelming - approximately 11 times fewer followers. This marked discrepancy suggests either a communication strategy predominantly oriented toward other social platforms, or content less adapted to the specificity and expectations of the Instagram audience.

Comparative analysis of likes on Instagram: Nicușor Dan vs. George Simion 

George Simion had no posts, only accepted collaborations on Instagram.

Nicușor Dan: Dominant performance with significant fluctuations Nicușor Dan demonstrates an extremely active presence on Instagram, with a considerably higher number of likes than his political rival throughout the analyzed period. The graph highlights several key points in his activity:

• A first important peak on April 29, when he recorded 118,708 likes
• Constant growth in the first days of May
• A spectacular maximum on May 4-5, reaching approximately 235,000 likes
• After a temporary decrease, a new significant peak around May 8 (approximately 140,000 likes)
• The end of the period shows an ascending tendency, with values around 100,000 likes This dynamic suggests a well-planned content strategy, with impact posts concentrated in strategic moments, generating consistent waves of engagement from his substantial digital community.

George Simion: no posts - only collaborations In strong contrast, George Simion records almost imperceptible values on the graph, with a line that remains constant in proximity to the zero axis. The only moments when small positive variations are observed are around the dates of May 4-5 and May 7 when he accepted collaboration posts with other people.

Comments on Instagram

Nicușor Dan: Clear domination and sustained community engagement Nicușor Dan records an impressive performance in terms of comments on Instagram, with values that fluctuate but remain constantly high throughout the analyzed period. The graph highlights a spectacular peak around the date of May 4-5, when the number of comments reaches approximately 2,600, followed by another significant moment on May 9 (about 1,700 comments). This dynamic reflects his ability to generate not just likes, but also active conversations around his content, demonstrating a high level of engagement from his substantial digital community.

George Simion: Minimal presence with a single moment of impact George Simion's activity in terms of comments on Instagram is almost non-existent for most of the analyzed interval, with a single notable exception around the date of May 5, when he records a peak of approximately 500 comments. This isolated moment of engagement, corroborated with the lack of own posts and preference for collaborations, suggests a deliberate strategy of concentrating resources on other social platforms where his audience is more receptive, such as TikTok.

TikTok

The graph below shows the number of followers on TikTok for the two presidential candidates George Simion and Nicușor Dan in three different time periods: April 23, May 2, and May 12.

Number of followers

George Simion: Consolidated force on TikTok George Simion stands out through his massive presence on the TikTok platform, with an impressive base of 1.3 million followers that remained stable in the first two analyzed periods. His digital influence continues to expand, recently reaching the threshold of 1.5 million fans - a significant increase of 200,000 new followers (15%). This performance confirms his dominant position in the political space on TikTok, where he has managed to build and maintain a substantial digital community.

Nicușor Dan: Accelerated growth from a modest base Nicușor Dan, although starting from a much more discrete position, records a remarkable dynamic of his audience. The evolution of his number of followers - from 188,000 (April 23) to 231,010 (May 2) and then to 327,400 in the most recent measurement - illustrates an accelerated ascending trajectory. Even if his digital community remains more than four times smaller than that of George Simion, the proportionally faster pace of growth (with successive jumps of 23% and respectively 42%) suggests an efficiency increase of his communication strategy on the platform and a potential for gradually reducing the gap in the following period.

Views trajectory

George Simion At the beginning of the analyzed period (April 24, 2025), George Simion recorded very reduced views, almost non-existent. The first days show minimal activity on the TikTok platform. But starting with April 27-28, his content begins to gain traction, reaching approximately 1 million views.

After a short period of moderate fluctuations, a dramatic increase is observed around May 3-4, culminating with a spectacular peak of approximately 10 million views on May 5.

After this remarkable peak, Simion's views decrease considerably, but remain at a significant level (3-4 million) compared to the beginning of the period. A second important peak appears on May 9, when views reach approximately 6 million, followed by a new decrease toward the end of the analyzed period.

Nicușor Dan Nicușor Dan begins the analyzed period with approximately 2 million views on April 24, positioning himself above George Simion. But in the following days, his views decrease dramatically, approaching 100,000.

Until the beginning of May, his activity on TikTok remains minimal, with very reduced views. Starting with May 3, a gradual increase is observed, which accelerates rapidly until May 5, when he reaches an impressive peak of approximately 6 million views - significant, but still less than Simion's maximum.

In the following period, Nicușor Dan's views follow a pattern similar to that of Simion, but at a slightly lower level, with the exception of the end of the period (May 11), when the two reach similar values of approximately 2.5 million views.

Remarkable is the fact that, although started from a larger initial base, Nicușor Dan went through a period of reduced activity before regaining his audience, suggesting a possible reevaluation of the content strategy at the beginning of May.

Analysis of TikTok likes for George Simion and Nicușor Dan

George Simion - Evolution of likes George Simion begins the analyzed period with a very reduced number of likes on TikTok around April 24, 2025. In the first days, his activity records modest values, below 20,000 likes.

Starting with April 27, a notable increase is observed, when his likes rise from 350,000 to over one million.

The turning point appears on May 4-5, when a spectacular increase is recorded, culminating with an impressive peak of approximately 10,000,000 likes on May 5. This explosion of popularity coincides with the peak of views observed earlier and indicates content extremely well received by the public.

After this maximum, likes decrease gradually, stabilizing around the value of 2,500,000-3,500,000, with a new significant peak on May 10 (approximately 6,000,000 likes). The end of the analyzed period shows a convergence with Nicușor Dan's values, around 2,500,000 likes.

Nicușor Dan - Evolution of likes Nicușor Dan begins the period with very low values of likes, almost zero in the first days of the analyzed interval. Until the beginning of May, his activity on TikTok generates a slow but constant increase, still maintaining below 200,000 likes.

A significant change occurs starting with May 3, when likes begin to increase rapidly, reaching a peak of over 6,000,000 on May 5. Although impressive, this maximum remains at less than half of the peak recorded by George Simion on the same day.

In the following period, the trajectory of Nicușor Dan's likes follows a pattern similar to that of Simion, but at lower values until May 9-10, when he reaches approximately 4,500,000 likes. At the end of the analyzed period (May 11), both politicians record similar values, around 2,500,000 likes.

It is noteworthy that, although started with a significant disadvantage in the first part of the period, Nicușor Dan manages to reduce the difference from George Simion toward the end of the analyzed interval, suggesting an increase in the efficiency of his content strategy on TikTok.

Analysis of TikTok comments for George Simion and Nicușor Dan

George Simion - Evolution of comments George Simion begins the analyzed period with a very reduced number of comments on TikTok around April 24, 2025, almost non-existent - 63 comments on April 24. In the first days, only minimal activity in the comments section is observed.

Starting with April 27, a sudden increase appears, reaching a first significant peak of approximately 17,000-18,000 comments on April 28. This first wave of engagement suggests that he published content that stirred debates or strong reactions from the public.

After a temporary decrease, comments fluctuate between 5,000 and 12,000 until May 3, when an accelerated increase begins, culminating with an impressive peak of approximately 34,000 comments on May 5. This maximum perfectly coincides with the peaks observed earlier in views and likes, confirming the impact of the content published on that day.

After this maximum point, comments decrease, but remain at a high level (15,000-22,000), with a new important peak on May 10 (approximately 32,000 comments). The end of the period shows a stabilization around the value of 20,000 comments, significantly higher than at the beginning of the analyzed period.

Nicușor Dan - Evolution of comments Nicușor Dan presents a much more modest evolution of comments compared to George Simion throughout the analyzed period. In the first half of the interval (April 24 - May 2), activity in his comments section is almost non-existent, with very low values, below 1000 and slightly above.

Starting with May 3, a gradual increase is observed, which continues until the end of the analyzed period. However, even at its maximum, the number of comments for Nicușor Dan does not exceed 8000 (recorded on May 9), representing a quarter of the peak recorded by George Simion.

The graph explicitly shows that on May 11, 2025, Nicușor Dan recorded 7,469 comments, which marks his best result in the analyzed period, but remains significantly below George Simion's performance.

Content distribution: George Simion vs. Nicușor Dan on TikTok

George Simion: Large fluctuations and impressive peak George Simion presents a dynamic evolution of shares on TikTok, marked by significant variations throughout the analyzed period. Starting with almost non-existent values on April 24, his account records two intermediate peaks around the dates of April 28 (approximately 12,000 shares) and May 1 (approximately 13,000 shares). The culminating point of his performance is recorded on May 5, when he reaches approximately 27,000 shares - coinciding with the peaks observed for other engagement metrics.

Nicușor Dan: Spectacular growth and maintenance at high levels Nicușor Dan demonstrates a surprising evolution in terms of content shares. Starting from a modest base (approximately 2,000) at the beginning, followed by a period of minimal activity, his performance radically transforms starting with May 3. His trajectory culminates on May 5 with approximately 27,000 shares - practically equal to George Simion. Remarkable is the fact that, unlike Simion, Nicușor Dan manages to maintain high values of shares even after this peak (between 17,000-23,000), with a consistent performance even after May 10. This represents the only metric in which Nicușor Dan not only catches up but even surpasses George Simion's performance in the second half of the analyzed period, suggesting content with greater potential for virality and relevance for his audience.

Democracy in the digital era: when reality is algorithmically distorted

The extended analysis of the online activity of presidential candidates confirms initial suspicions of coordinated inauthentic behavior in the Romanian digital ecosystem. The documented patterns - from impossibly synchronized engagement peaks to networks of suspicious accounts and abnormal interaction reports - suggest deliberate manipulation of political discourse.

What is more worrying is the convergence of strategies in the second part of the campaign. After public denunciations by both candidates regarding "attacks" with fake accounts, paradoxically, an intensification of these practices is observed, not a diminishment. The difference is not in abandoning problematic tactics, but in adopting them on a larger and more sophisticated scale. Moreover, the present analysis reveals strategies of inauthentic coordination only in the context of the official accounts of candidates. The reality of social networking platforms indicates a much deeper inauthentic void, with hundreds of accounts that satirize, denigrate, and raise candidates in algorithms through coordinated actions, accounts that appeared overnight and loading the digital information space with discussions that, in reality, are not facilitated by real users.

In the digital era, popularity can be fabricated, not just earned - and the fingerprint of this fabrication remains visible in data. As Romania heads toward a decisive moment, voters deserve transparency regarding the possibility of an online impulse that is not stimulated by a conscious confirmation bias, but possibly by an artificially created consensus. Without improving the detection and discouragement of coordinated inauthentic behavior, social platforms risk becoming not an alert digital public marketplace, but an algorithmized bazaar where political visibility is available to those with resources to purchase it. Protecting citizens and, implicitly, democracy itself, requires accountability from all actors in the information ecosystem - from very large online platforms to institutions responsible for regulation - to guarantee that electoral preferences are formed based on authentic debates, not algorithmic illusions.