**MEDIA ECOSYSTEM OBSERVATORY BRIEF**
# **Power shift: The rise of political influencers in Canada**
November 2025
# **Contents**
| Executive summary | 1 |
|-------------------|----|
| Résumé exécutif | 1 |
| Context | 4 |
| Questions | 4 |
| Conclusions | 13 |
| Methodology13 | |
| Bibliography | 15 |
| Endnotes | 15 |
| Contributors | 16 |
# **Power shift: The rise of political influencers in Canada**
# **Executive summary**
The line between lifestyle, commerce, and news has never been blurrier, and the influencer is the figure traversing all three. Recent years have seen the dramatic rise of the *political influencer* people or groups that primarily produce political/ current-affairs/social-discourse content and have amassed a following doing so. This potent, digitally native group is now leveraging its reach to shape public debate. In the B.C. 2024 election1 , we found that while news outlets posted the most, political influencers generated 63.9% of all engagement. In the 2025 federal election analysis2 , they accounted for almost half (47%) of all political content posted in our dataset, compared with 28% from news outlets and 18% from politicians. Beyond sheer volume, influencers attract the bulk of attention and bridge media, campaigns, and the public. Influencers have become the information brokers of the internet, setting the pace for political conversation that traditional political parties, media outlets, and advocacy organizations struggle to match.
This information ecosystem brief maps the role of political influencers in Canada's online information ecosystem from January 2024 – July 2025.
# **Changement de pouvoir : l'ascension des influenceurs politiques au Canada**
# **Résumé exécutif**
La frontière entre style de vie, commerce et information n'a jamais été aussi floue et la figure de l'influenceur navigue désormais entre ces trois univers. Ces dernières années ont vu l'ascension spectaculaire des **influenceurs politiques** : des individus ou des groupes qui produisent principalement du contenu sur la politique, l'actualité et les débats sociaux, et qui ont bâti une audience autour de cela. Ce groupe, à la fois puissant et natif du numérique, met aujourd'hui à profit sa portée pour façonner le débat public. Lors de l'élection en Colombie-Britannique en 2024, nous avons constaté que, bien que les médias d'information publiaient le plus de contenu, les influenceurs politiques avaient généré 63,9 % de l'ensemble de l'engagement. Dans notre analyse de l'élection fédérale de 2025, ils représentaient près de la moitié (47 %) de tout le contenu politique publié dans notre ensemble de données, contre 28 % pour les médias d'information et 18 % pour les politiciens.
Au-delà du volume, les influenceurs captent l'attention, relient les médias, les campagnes et le public. Ils sont devenus les nouveaux courtiers de l'information, imposant le rythme du débat politique en ligne que les partis politiques, les médias traditionnels et les groupes de défense ont du mal à suivre.
Ce portrait de l'écosystème de l'information cartographie le rôle des influenceurs politiques dans l'espace numérique canadien, We use four sources of evidence:
- 1. A nationally representative survey to see how Canadians view and engage with influencers;
- 2. A data-driven grouping of influencer accounts across major platforms to see who resembles whom;
- 3. An automated topic analysis to compare which issues influencers drive versus the issues driven by other actors;
- 4. A time-series on posting and engagement to track how much they post and how much attention they get.
# **Key takeaways**
- E **Influencers are now central to Canadians' information diets:** 42% of Canadians report engaging with influencers on political topics. This number shoots up for younger Canadians (64%) and partisans (49%), meaning online conversations about politics are increasingly shaped by the content influencers produce.
- E **Influencers command more attention online than traditional institutions:** They have significantly more reach than either news media or politician and party accounts in Canada. Their activity represents approximately 51% of all posts and 62% of the engagement across accounts tracked.
- E **Canada's political influencer ecosystem is fragmented but vibrant:** Network analysis shows a highly clustered structure of political influencers across six distinct communities. Each community reflects different leanings, illustrating both diversity and polarization within the political influencer sphere.
- E **Influencers are highly responsive to political cycles and play an active role in amplifying or framing key events:** They were most active during major elections including the 2024 B.C. and American elections and the 2025 Canadian federal election, with peak post volume reaching 220% above the minimum level of activity. They are highly responsive to emergent political stories and capture the majority of attention on mainstream policy topics like healthcare, education, cost of living, and U.S.-Canada relations.
de janvier 2024 à juillet 2025. Quatre sources de données soutiennent notre analyse :
- 1. Un sondage national représentatif pour comprendre comment les Canadiens perçoivent et interagissent avec les influenceurs;
- 2. Une analyse de regroupement de comptes d'influenceurs sur les principales plateformes, afin d'identifier les similarités et les communautés;
- 3. Une analyse automatisée des sujets pour comparer les thèmes mis de l'avant par les influenceurs à ceux portés par d'autres acteurs;
- 4. Une analyse temporelle des publications et de l'engagement pour suivre leurs niveaux d'activité et de visibilité au fil du temps.
# **Principaux constats**
- E **Les influenceurs sont désormais au cœur des habitudes d'information des Canadiens** : 42 % des Canadiens déclarent interagir avec des influenceurs sur des sujets politiques. Ce chiffre grimpe à 64 % chez les jeunes et à 49 % chez les partisans, indiquant que les discussions politiques en ligne sont de plus en plus façonnées par le contenu produit par ces acteurs.
- E **Les influenceurs attirent davantage l'attention en ligne que les institutions traditionnelles :** leur portée dépasse largement celle des médias ou des comptes de politiciens et de partis. Ils représentent environ 51 % de toutes les publications et 62 % de l'engagement total parmi les comptes suivis.
- E **L'écosystème des influenceurs politiques au Canada est fragmenté mais dynamique :** l'analyse des réseaux révèle une structure très compartimentée, répartie en six communautés distinctes. Chacune reflète des sensibilités différentes, illustrant à la fois la diversité et la polarisation du milieu.
- E **Les influenceurs réagissent fortement aux cycles politiques et participent activement à l'amplification ou à la mise en récit des grands événements** : leur activité atteint un pic lors des principales élections, notamment celles de la Colombie-Britannique et des États-Unis en 2024, ainsi que l'élection fédérale canadienne de 2025, avec des volumes de publication jusqu'à 220 % au-dessus des niveaux minimaux. Ils réagissent rapidement aux nouvelles politiques émergentes et captent la majorité de l'attention sur des thèmes majeurs comme la santé, l'éducation, le coût de la vie et les relations canado-américaines.
E **U.S. politics is perennially important for Canadian discourse:** It remains a steady presence across content types, with media outlets allocating the largest share of their posts (27.4%) to U.S.-related topics. Influencers follow below media (23.9%), reflecting somewhat lower but still significant engagement with U.S. issues.
# **Why it matters**
Influence is shifting from institutions to individuals with news outlets and political parties and advocacy groups not reaching as many people as they used to. As political influencers rise in prominence, they are not subject to formal accountability mechanisms. Effective strategies for public communication, audience reach, mobilization, and counter-misinformation require community-level targeting, partnerships with bridge accounts, timing to event windows, and metrics that track who is heard by whom, not just who posts most.
E **La politique américaine continue de jouer un rôle majeur dans le débat public canadien :** elle occupe une place constante dans les contenus publiés. Les médias y consacrent la plus grande part de leurs publications (27,4 %), suivis des influenceurs (23,9 %), témoignant d'un intérêt légèrement moindre mais toujours significatif pour les enjeux américains.
## **Pourquoi c'est important**
L'influence se déplace des institutions vers les individus. Les médias d'information, les partis politiques et les groupes de défense n'atteignent plus le même public qu'auparavant. Or, à mesure que les influenceurs politiques gagnent en visibilité, ils ne sont pas soumis aux mêmes mécanismes de reddition de comptes formels.
Les stratégies efficaces en matière de communication publique, de mobilisation et de lutte contre la désinformation doivent désormais intégrer une approche centrée sur les communautés, des partenariats avec des comptes "ponts", une synchronisation avec les moments clés du calendrier politique et des indicateurs d'écoute, qui mesurent qui entend qui, plutôt que seulement qui publie le plus.
Andrew Neel/Unsplash
# **Context**
The digital landscape has fundamentally reshaped how political information is created, circulated, and consumed. Our research examines a distinct and increasingly influential group that we define as *political influencers*. These are individuals or groups who consistently produce political content and have the capacity to shape public opinion in the Canadian information ecosystem. In this report, *political content* refers to topics with clear political relevance, like elections, public policy, international conflicts, human rights, or social justice.
Political influencers are heterogeneous: they include activists, independent commentators, journalists operating in hybrid roles, partisan communicators, and community organizers. Political influencers operate across platforms (e.g., X, TikTok, YouTube) and formats, often building durable communities rather than one-off reach spikes.
This report situates political influencers within Canada's online information ecosystem, mapping who they are, how active they are, what topics they drive, and how their engagement compares with news outlets and other actor types. Our goal is to provide an evidence-based view and understanding of their role and influence.
# **Questions**
This report focuses on five practical questions that structure the analyses:
#### E **Why should we care about political influencers?**
Using a nationally representative survey, we examine how people view influencers' credibility, usefulness, and whether they are seen as reliable information sources or mainly as entertainers.
# E **Who are these political influencers?**
We define *political influencers* using a clear, data-driven rule: they consistently post about politics/current affairs and generate measurable audience impact. This lets us separate them from traditional media and politicians. We then cluster accounts by their content to surface distinct clusters and the issues that bind them.
#### E **How active are they?**
We compare posting frequency and cadence to news outlets and other actor types, highlighting event-driven spikes and simple platform patterns. This helps gauge their role as key content creators, especially around elections.
# E **What issues do political influencers focus on?**
We use topic modeling to extract topics and present a highlevel policy distribution (e.g., economy, housing, healthcare). Running the model on influencer posts and on the overall corpus lets us compare shares and relative lift to see which issues are driven or shaped by influencers. We also run a parallel check for U.S.-related topics: we measure how often influencers talk about them versus the overall corpus.
E **How does the discussion compare between influencers and other actors ?** We compare influencers' share of posts with their share of attention (likes) across topics to see where they over- or under-perform. We then map their position in the network relative to legacy media and institutions, estimate effective reach, and check how much their content stays in silos versus spills across communities.
To answer those questions, we combine survey and computational evidence. The survey component draws on a nationally representative poll that measures respondents' perceptions of
**Figure 1 Self-reported engagement with influencers**
and engagement with political influencers. On the computational side, we analyze data from six social platforms between January 2024 and July 2025 spanning ~9M posts in total, of which ~4M are by political influencers.
# **Question 1 Why should we care about political influencers?**
Influencers now play an important role in the media environment, with a substantial share of Canadians reporting regular exposure to their content. To better understand this phenomenon, we surveyed 1,431 Canadians between August 28 and September 3, 2025, using a nationally representative sample by age, gender, and region, to examine their engagement with and attitudes toward online influencers.
As shown in Figure 1, 42% of Canadians engage at least sometimes with political content from influencers. If you broaden the question to all types of content, not just political, then this number grows to over half of Canadians engaging with influencers (54%). We use "engagement" broadly to refer to activities such as consuming, liking, sharing, or commenting on posts.
There is significant variation by group. Nearly two thirds (64%) of Canadians aged 18 to 34 engage with politically-oriented influencers but that number drops to only 32% for Canadians over 55. Engagement with political influencers strongly correlates with partisanship, with those identifying with a political party 20% more likely to engage with influencers than those who do not. Interestingly, we do not see significant variation by gender or across political parties.
Overall, this speaks to the growing power that influencers occupy in our public discourse. Our partisan politics, increasingly contested in the digital public square by younger voices, is driven by those who report substantially higher engagement with political influencers.
Our survey also included a question about why Canadians engaged with influencer content, with the results shown in Figure 2. Respondents most often cited staying informed about current events (41%) and learning about topics of interest (40%). Influencers have become an important source of news, information, entertainment, and commentary for millions of Canadians.
We further investigated how Canadians came to discover influencers, with the results in Figure 3. The two most significant ways people discover influencers are through their friends' recommendations (32%) and platform algorithms (32%). Discovery through direct search remains relatively low at just 20%. Influencer discovery is shaped by the structural dynamics of social media platforms, with only a smaller share of respondents (16%) reported encountering political influencers through other means, including traditional media such as television, radio, newspapers, and podcasts, or through more incidental interactions online and offline.
**Figure 2 Reasons for engaging with political influencers**
**Figure 3 How respondents first encountered political influencers**
# **Question 2 Who are these political influencers?**
"Influencer" can mean many things: celebrity creators, meme pages, activists, journalists etc. Our survey confirms this ambiguity: people apply the label inconsistently and frequently cited news media outlets in an open-ended question. For the analysis using social media data, we use an operational, data-driven definition: a political influencer is a person or group who consistently produces political content and has the power to shape public opinion.
To understand political influencers, we collected public posts (n = 4,103,984) across social media platforms from individually-identified influencers (n = 1,097) from January 2024 to July 2025. See the methodology section for additional information on the identification of influencers and the collection of data.
#### **Influencers categorization**
We begin with a categorization of Canadian political influencers. We employ a content-similarity and social-network approach where the semantic similarity of all posts from each influencer is compared to that of each other influencer. We then apply a clustering algorithm designed to find tightly knit groups (or "communities") of influencers who consistently share similar topics. The result was a modularity score3 of 0.65, which is a strong indicator that these groups are genuine, significant, and not just random collections of accounts. We then manually reviewed and labeled these algorithmic clusters (using account metadata, public posts, and central nodes or exemplars) to effectively validate the clustering algorithm.
**Figure 4 Influencer categorization map**
This process yielded six main clusters, each with a distinct topical focus and communication style:
**Political commentariat:** The largest and most internally diverse cluster brings together activists, protest movements, journalists, organizations, and policy experts across the political spectrum. Unlike more focused clusters, it reflects the broad Canadian political arena where debates on core national issues take [place.](http://place.in) What unites influencers in this community is not a shared ideological position or policy focus, but rather their consistent engagement with any and all issues that are currently trending in the Canadian digital sphere, from major political controversies to breaking news and policy debates. Examples include Andrew Coyne, BertaProudDad, Bruce Anderson, Chantal Hébert, Keean Bexte, Lucie Laurier, and Mario Zelaya.
**Identity-centric debate:** A community with opinion-driven content centered on culture and identity debates. This includes opinion posts about Israel and Gaza, the Russian invasion of Ukraine, debates over immigration and Canadian multiculturalism, as well as posts related to gender and race. This community also captures politically diverse influencers across the left-right political spectrum. Examples include Elizabeth Renzetti, Fae Johnston, Nora Loreto, and Warren Kinsella.
**Culture and community:** This cluster is characterized by culture-first influencers including indigenous creators, prominent cultural figures, and others featuring Canadian events, communities, and places. Common themes within this cluster include multicultural festivals and arts, sports, and indigenous identity. Examples include Ian Hanomansing, Rick Mercer, and Shina Nova.
**Pop politics:** A cluster blending pop-culture commentary, sports, humor, and lifestyle voices with a light focus on public policy issues such as healthcare and carbon tax. More entertainmentand personality-driven compared to the "Culture and community" cluster, with a focus on soft news and cultural discourse rather than hard politics. Examples include Liz Plank, Gad Saad, Jody Vance, and Quick Dick McDick.
**International humanitarian politics:** A smaller, advocacy-driven cluster of international human rights and decolonial groups, activists, academics, lawyers, and journalists. Common topics and themes typically center around humanitarian concerns and international affairs, such as refugees and asylum, ICC and international law, and the role of bodies like NATO. Examples include Pam Palmater, Shree Paradkar, and Yves Engler.
**Alternative media politics:** A small but distinct cluster tied to anti-establishment, sensational, and sometimes conspirator-
**Figure 5 Monthly post volume across platforms (January 2024– July 2025)**
ial narratives, overlapping with U.S. far-right online spaces. Common themes include vaccine conspiracies, election fraud, and general institutional distrust. Examples include Freedom News, Veterans for Freedom, and Mark Friesen.
By comparing the content of what these influencers talk about we can better understand how they engage with one another. Far from a strict left-right or partisan divide, we find that influencer clusters are more keenly focused on distinct topics and events of interest. In this way, influencers occupy a central space connected to but distinct from the logic of partisan politics.
# **Question 3 How active are influencers?**
Influencers are highly active and engaged on social media. We tracked the activities of 1,097 influencers active over the period examined here who collectively produced 4.1 million posts and garnered 815.7 million engagements, as measured by likes. This far exceeds that of politicians (1.1 million posts and 184.0 million engagements) or media outlets (2.8 million posts and 312.6 million engagements) over that same time period. Influencers represented 51% of all posts and 62% of all engagements across the time period examined here. For the average Canadian who uses one or more social media platforms to get political information, a significant or even majority of the content they see comes directly from influencers.
However, the activity of influencers is highly variable, with their posting frequency and engagement (average likes received) levels fluctuating sharply in response to major political events. During our period of analysis from January 2024 to July 2025, post activity peaked at 220% of its lowest point. Influencer activity, particularly post volume, increased significantly around election periods.
**The election effect:** The analysis of post volume shows two distinct peaks for influencers, which strongly correlate with significant political events. The first and highest peak occurred in October and November of 2024, a period that encompassed the British Columbia provincial election on October 19 and the United States presidential election on November 5. Our topic modeling confirms that this surge in activity was directly tied to discussions around these two electoral events.
The second major peak was in April 2025, which coincided with the Canadian federal election. Political cycles serve as powerful catalysts for online discourse, with influencers leveraging these moments to create a high volume of content. A comparison with politicians' activity reveals a similar pattern, with their post volume also peaking around these times.
Influencers, despite having far fewer resources than either politicians or news outlets, produce more posts and garner greater engagement than either of these other groups. Their activities, however, are closely tied to political event cycles rather than following a simple upward or downward trend.
# **Question 4 What issues do influencers focus on?**
To identify what political influencers talk about, we applied topic modeling to their ~4M public posts. We focus only on influencers to isolate their attention and avoid confounding other actors' posting habits or volume differences. We used BertTopic and a clustering algorithm to identify a broad range of themes and then, similar to the influencer cluster analysis, manually reviewed exemplar posts and validated and labelled the topics. Broadly, the topics fall into three main themes.
First, Canadian partisan politics; domestic policies; and international events. Influencers largely focus on discussions related to Canadian politics, with a strong emphasis on the 2025 federal election. Tightly-knit discourse covers updates and assessments of the Liberal and Conservative Parties and their respective leaders, framed around government policy and public trust.
Second, key domestic policy issues frequently discussed include the state of the Canadian economy, the cost of living, housing, healthcare, and increase in drug use and crime.
And third, influencers also had a high volume of discussion of international events that often but not always are tied to Canadian foreign policy and interest, such as the war in Ukraine, the Israel-Palestine conflict, and the Trump administration.
The wide diversity of topics and themes reflects the diversity of the Canadian influencer community. Unlike the United States, where culture war issues anchor much of the elite influencer conversation (CITE), we find a pronounced spread of ideas.
While classifying topics we did identify a large number of US-focused discussions and given significant concerns about U.S. influence and interference in the Canadian information environment, we assessed the extent to which Canadian influencers had a strong U.S. orientation.
#### **U.S. vs. non-U.S. related topics**
Across all topics, 86 topics were U.S.-related vs. 390 non-U.S. related topics. However, when we look at the top 50 most engaging posts, we see a much higher proportion of U.S.-related content, with 17 out of 50 being U.S.-related topics. This includes topics such as:
**Healthcare policies:** the impact of overturning Roe v. Wade on women's healthcare, access to medical care, controversy around COVID-19 policies and vaccines and RFK Jr.'s appointment as U.S. Secretary of Health and Human Services, and comparisons of U.S. and Canadian health services
**US foreign policy:** The Trump administration's response to the war in Ukraine and withdrawal from global cooperation such as the Paris Agreement and the WHO
**Trump administration:** Criticisms of Trump's administration, including his relationship with Elon Musk and the operation of federal institutions such as the Treasury Department, FBI, and National Guard
**US-Canada relations:** Reactions and criticism to Trump claiming Canada as the 51st state and subsequent calls to to stop purchasing American products
The Canadian media ecosystem is intertwined with American politics. U.S. political events mobilize a significant amount of online discourse in Canada, shaping conversations around healthcare policies, foreign policy, and international stability. Figure 6 shows the weekly number of U.S.-related posts and the percentage of all posts that referenced the U.S. by Canadian influencers from January 2024 to July 2025.
**Figure 6 Weekly volume of posts on U.S.-related topics (January 2024– July 2025)**
**Figure 7 Participation vs. attention by topic (account-normalized)4**
From January 2024 to July 2025, Canadian discourse related to American politics largely centered around Donald Trump and his presidential campaign. Increased content activity is related to two assassination attempts during his rallies in July and September 2024, his election in November 2024, and the inauguration in January 2025. His suggestion of Canada as the 51st state and the beginning of tariff discussions in early January 2025 also sparked significant conversation. Canadian influencers are keenly aware and active around key political events in the United States and their implications for Canada. While overall volume declined in mid-2025, the share of U.S.-related content remained relatively high, suggesting sustained interest despite lower posting activity.
# **Question 5**
# **How does the discussion compare between influencers and other actors ?**
Building on Q4's influencer-only model (~4M posts), we add a full-corpus topic map (~9M posts across influencers, politicians, news, and institutions) to place influencer themes in context. This lets us compare what each group talks about, quantify overlap vs. unique topics, and see where influencers lead, mirror, or diverge.
Influencers cluster at the partisan core of the overall political discussion (Liberal–Conservative rivalry), led conversation around the 2024 BC election but not the 2025 Ontario race,
**Figure 8 Content-similarity network of Canadian actors5**
but are largely absent from local politics discussion which is dominated by news outlets. International flashpoints including Gaza-Israel, Ukraine-Russia, and Trump are major subjects of conversation across all three groups. Economic and social issues (GDP, debt, healthcare, Indigenous rights, gender) sit adjacent to this core and often blend into culture-war debates.
In Figure 7, we compare participation (post share) and attention (like share) by topic across all actor types. The left panel shows the composition of discourse by post share and the right panel by like share. The dashed line in each panel marks the unweighted average influencer share across topics. The posts are ordered descending by what percentage of all posts in that topic are made by influencers.
High influencer engagement is strongest on commentary-heavy themes. Topics include call-outs and mockery, media and journalism criticism, and Trump and US-Canada relations. For these topics, influencers both post the most and receive the most engagement. By contrast, public incident and service topics such as weather events, wildfires & evacuations, crime and public safety, community support show greater posting shares for institutional actors (news outlets, civil society organizations, and government); even in these spaces, however, the engagement panel typically tilts back toward influencers, indicating that audiences still reward commentary layered on top of reported events.
The panels also show a clear asymmetry: influencers constitute roughly a third of the typical conversation across topics, yet they capture a majority of overall engagement. In other words, influencers attract disproportionately more audience attention as compared to politicians and news outlets. This gap is not driven by a single topic and appears broadly.
To map the online information ecosystem's internal structure, as shown in Figure 8, we followed the same methodology described in Q2 and shown in Figure 4. This time, however, we applied it to the entire dataset, encompassing all entities: influencers, politicians, news outlets, and institutions. This lets us map the entire ecosystem and see who clusters together versus who stands apart. The network is sparse and fragmented, with tight clusters and few links between them. Actors across the system tend to self-organize into tightly-knit communities based on shared interests and identities (modularity score of 0.75). This high degree of fragmentation (distance between clusters) with tight clustering suggests a structure where discussions and narratives flow rapidly and intensely within these individual
communities, while connections and the spread of information between these distinct groups are sparse. This fragmented structure is consistent with an ecosystem where users likely experience reinforcement of their views, raising challenges for efforts aimed at broad public communication. This finding is crucial for understanding influencer behavior:
**Siloed content**: The large blue cluster of influencers confirms that these actors are highly effective at forming a singular discourse community. They largely talk about the same subjects, use similar framing, and share content that is thematically highly consistent (high content similarity) with each other.
**Concentrated power**: This extreme internal coherence means that when influencers post content, the high volume of external public attention (likes, shares, comments) they generate is overwhelmingly concentrated within these segregated ideological pipelines. The network structure thus explains how influencers manage to achieve their outsized share of engagement: by operating within highly specialized and reinforcing content silos that minimize cross-cutting exposure while maximizing internal resonance.
This network architecture confirms a critical reality of Canadian political communication: siloed discourse that encourages actors to remain in segmented, reinforcing information space.
# **Conclusions**
The Canadian political information ecosystem is no longer a centralized arena dominated by newsrooms and party institutions and actors. It has become a decentralized network where political influencers play a persistent, central, and highly visible role. Our data shows that political influencers capture a disproportionate share of attention and engagement across many topics, from domestic politics to international conflicts. This influence relies less on rare viral spikes and more on steady, high-cadence and high-engagement posting.
Network structure helps explain these outcomes. The network underpinning the Canadian information ecosystem is fragmented and siloed, with key actors forming distinct content communities.
Taken together, these findings point to a mature but stratified political influencer sphere. They are durable agenda setters in commentary-heavy spaces and effective amplifiers even on incident-driven news, where audiences reward interpretation layered on reporting. For communicators, policymakers, and civil society, the implication is clear: broadcast-style messaging is insufficient. Effective strategies must work with the grain of the ecosystem, targeting specific communities, partnering with credible bridge accounts, timing outreach to event windows, and measuring not only who posts, but who is heard by whom.
# **Methodology**
# **Dataset**
Social media data cover January 2024–July 2025 across X, Instagram, TikTok, YouTube, and Bluesky. The data collection is based on a comprehensive seed list of entities relevant to the Canadian media ecosystem (see Pehlivan et al., 2025). This list includes politicians, news outlets, and influencers. A seed entity must be active on at least one platform and is defined as a person, group, organization, or media product of substantive interest to the ecosystem. In total, the seed list comprises 4,476 unique entities including 1,193 political influencers. Data was collected using various methods, including official APIs for platforms like Bluesky and YouTube, and custom web scraping for platforms like X, Instagram and TikTok. In total, the dataset includes over 4 103 984 posts from political influencers and 9 842 368 posts from all tracked entities combined. However, creating a fully comprehensive dataset is not possible due to the dynamic and constantly evolving nature of social media; nonetheless, this dataset provides a representative snapshot of the Canadian information ecosystem.
# **Influencer categorization network**
We clustered Canadian influencers using their public posts from January 2024 to July 2025 across X, Instagram, YouTube, TikTok, and Bluesky. For each account, we computed a mean text embedding over the period and built an influencer–influencer similarity graph using cosine similarity; we added an edge when similarity > 0.80. Communities were detected with a modularity-maximizing (Louvain) algorithm in Gephi; the resulting network shows high community structure (modularity = 0.65). For visualization only, we applied ForceAtlas2 in Gephi to position nodes so that densely connected groups appear closer in the plane. Node size calculated by using Page
Rank algorithm. Community labels were assigned after a brief qualitative review of top accounts per cluster.
## **Canadian media ecosystem network analysis**
We repeated the same pipeline for the full ecosystem—including influencers, news outlets, politicians,government and civil society organizations, over January 2024 to July 2025. We computed a mean content embedding per account, measured pairwise cosine similarity, and added edges at > 0.80. We visualized the graph in Gephi (ForceAtlas2). In this view, node color encodes actor type (e.g., influencer, news outlet, politician etc.). This layout makes it easy to compare where each actor type sits, spot bridge accounts linking clusters, and see how institutional nodes and individual influencers co-locate or separate across communities.
# **Topic modeling**
To understand the key issues driving online conversations, we used BERTopic, a neural topic modeling technique. BERTopic is an advanced topic modeling technique that uses a combination of modern AI models and statistical methods to find and define topics within a collection of documents. Unlike older methods that rely on word counts, BERTopic understands the semantic meaning of text.
We performed two separate analyses. First, to categorize influencers, we applied BERTopic directly to all influencer posts. Second, to measure influencers' ability to shape the broader discourse, we ran a topic model on the entire dataset. To avoid missing important topics and to account for temporal shifts, we calculated topics in three-month quarters and aligned related themes over time.
Additionally, we used GPT-5 to classify each topic to determine if it was related to the United States. This helped us analyze the outsized influence of U.S. politics on the Canadian information ecosystem, particularly among highly engaging posts.
#### **Survey**
Between August 28 and September 3, 2025, we fielded six questions on the role of influencers in political activities as part of our monthly tracking survey, administered through Qualtrics. The questions were as follows:
*We want to ask you a few questions about social media influencers. By social media influencers, we mean accounts that mainly create content that can affect the decisions of others through their popularity or relationship with their audience.*
# **Q1. How often do you engage with content from influencers on the following topics?**
- E Political content
- E Non-political content
- E Not at all
- E Sometimes
- E Most of the time
- E All the time
### **Q2. [only shown to those that don't say not at all for each relevant category]**
How much do you trust influencers when they post about political content?
- E No trust at all
- E Very little trust
- E A fair amount of trust
- E A lot of trust
# **Q3. [only shown to those that don't say not at all for each relevant category]**
How much do you trust influencers when they post about non-political content?
- E No trust at all
- E Very little trust
- E A fair amount of trust
- E A lot of trust
#### **Q4. [only shown to those that don't say not at all for each relevant category]**
Thinking about influencers **who post about politics or current events**, who is the first person that comes to mind?
## **Q5. How did you first come across this influencer? (Select one)**
- E Recommended by the platform algorithm or feed
- E Shared or recommended by friends or others
- E Search results
- E Other (please specify)
## **Q6. Why do you engage with influencers on social media** *about political or current events***? (Select all that apply)**
- E To stay aware of current events
- E I care about the issues they post about and their viewpoints
- E They are influential on important topics
- E To learn about topics of interest
- E To stay connected with my community
- E They are entertaining
- E Other (please specify)
# **Endnotes**
- 1 https://meo.ca/work/british-columbia-election-informationecosystem-project
- 2 https://www.cdmrn.ca/publications/the-canadianinformationecosystem-during-the-2025-federal-election
- 3 Modularity is a measure used in network science to evaluate how well a division of a network captures densely connected groups with sparse connections between them. Intuitively, it compares the actual density of links within proposed communities to what would be expected if the same network were randomly connected. Higher values suggest a meaningful community structure rather than a random partition.
- 4 Values show the breakdown by actor type (influencer, politician, etc.) within each topic. For each topic × actor type, we count posts per account and likes per account, take the mean for that actor type, then rescale within the topic so shares sum to 100%. (Applied separately for posts and for likes.)
- 5 This graph visualizes the ecosystem structure. Each node is an account, with its color identifying the actor type (e.g., influencer, news outlet). Edges link accounts sharing similar content, and the overall layout algorithmically maps these relationships, allowing us to see how different actor types cluster or separate.
# **Bibliography**
Dubois, E. & Stahl, L. (2025). Influencers and Elections: The many roles content creators play. Pol Comm Tech Lab, University of Ottawa. polcommtech.com/research/influencers-and-elections
Pehlivan, Z., Park, S., Abrahams, A. S., Desblancs-Patel, M., Steel, B. D., & Bridgman, A. (2025). Building a Media Ecosystem Observatory from Scratch: Infrastructure, Methodology, and Insights. *arXiv preprint arXiv:2506.10942*.
Bastian, M., Heymann, S., & Jacomy, M. (2009). Gephi: An open source software for exploring and manipulating networks. Proceedings of the International AAAI Conference on Web and Social Media, 3(1), 361–362.
Blondel, V. D., Guillaume, J.-L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 2008(10), P10008.
Jacomy, M., Venturini, T., Heymann, S., & Bastian, M. (2014). ForceAtlas2, a continuous graph layout algorithm for handy network visualization designed for the Gephi software. PLOS ONE, 9(6), e98679.
Newman, M. E. J. (2006). Modularity and community structure in networks. Proceedings of the National Academy of Sciences, 103(23), 8577–8582.
Brin, S., & Page, L. (1998). The anatomy of a large-scale hypertextual Web search engine. Computer Networks and ISDN Systems, 30(1–7), 107–117. https://doi.org/10.1016/S0169-7552(98)00110-X
Grootendorst, M. (2022). BERTopic: Neural topic modeling with a classbased TF-IDF procedure. arXiv preprint arXiv:2203.05794.
# **Contributors**
#### **Zeynep Pehlivan, Ph.D.**
Zeynep leads the Digital Trace team at the Media Ecosystem Observatory. She is a computer scientist with over 15 years of experience applying advanced computational methods to social science research.
#### **Junyan Zhu, Ph.D.**
Junyan is the survey lead at the Media Ecosystem Observatory and a postdoctoral fellow in the Department of Political Science at the University of Toronto.
#### **Chris Ross**
Chris is a survey analyst for the Media Ecosystem Observatory. He has a MA in Political Science from McGill University.
#### **Diya Jiang**
Diya is a researcher at the Media Ecosystem Observatory (MEO). She specializes in the politics of international trade, public opinion, and US-Canada Relations.
# **Saewon Park**
Saewon is a senior data analyst at the Media Ecosystem Observatory (MEO). Prior to MEO, she worked as an analyst in the tech industry.
#### **Esli Chan**
Esli is a research fellow at the Media Ecosystem Observatory (MEO). She specializes in gender issues, media and technology, and radicalization and extremism.
## **Jennie Phillips, Ph.D.**
Jennie directs the Project on Information Ecosystem Resilience and leads information incident response. Her expertise is in disaster and emergency management, public education and resilience, and the digital transformation of society.
#### **Aengus Bridgman, Ph.D.**
Aengus is the Director of the Media Ecosystem Observatory and a leading expert in misinformation, digital activism, and the impact of online media on democratic life.
This brief was made possible by the considerable work of a large number of researchers in Canada. In no particular order, we would especially like to thank:
Elizabeth Dubois Olivia Melanson
Ashley Vu Nader Hedfi
Designed by Brian Morgan
The work of the Media Ecosystem Observatory and the Canadian Digital Media Research Network is funded in part by Heritage Canada's Digital Citizen Initiative.
