VIS x GenAI 2025

The chart reveals a stark difference in mean citation counts based on the presence of a Graphics Replicability Stamp. Papers with an "Unknown" stamp status have a significantly higher mean citation count (70.721) compared to those explicitly marked with a replicability stamp ("X"), which average only 4.778 citations. This suggests that papers without a clear replicability designation may be more widely cited, potentially indicating a bias toward established or less scrutinized works, or a lack of awareness of the replicability stamp's significance within the research community.

The chart reveals that the majority of stamps are distributed across three paper types: J, C, and M, with J having the highest count (1,972 stamps), followed by C (1,559 stamps) and M (319 stamps). Additionally, a smaller subset of stamps (27) labeled as "X" is exclusively associated with paper type J. This suggests that paper type J is not only the most prevalent overall but also uniquely supports stamps of type "X," highlighting its distinctiveness in the dataset.

The citation trends over the years reveal a fluctuating but generally declining pattern after peaking in the mid-2000s. Citations steadily increased from 1990, reaching a notable high in 2006, followed by a sharp drop in 2008. Although there were smaller peaks in 2011 and 2014, the overall trend since 2015 shows a consistent decline, with citations falling dramatically after 2018. This suggests a possible shift in research focus, reduced relevance of the data source, or changes in citation practices over time.

The chart reveals distinct trends in citation counts over time for three paper types (C, J, and M). Paper type C shows a peak in average citations during the late 1990s to early 2000s, followed by a sharp decline after 2013, with data gaps in recent years. Paper type J demonstrates a delayed but higher peak in citations, starting around 2006 and declining steadily after 2012, eventually reaching near-zero by 2024. Paper type M consistently has the lowest citation counts, with sporadic data and no significant upward trends, indicating limited impact or inconsistent publication activity. Overall, paper types C and J dominated citations historically, but both exhibit declining relevance in recent years.

The chart reveals that journal papers (J) constitute the majority of conference submissions, accounting for 51.6% of the total, followed by conference papers (C) at 40.2%. Miscellaneous papers (M) represent a small fraction, making up only 8.2% of submissions. This indicates a strong preference for journal papers in conferences, suggesting they may be perceived as more formal or impactful compared to other types, while miscellaneous papers are relatively rare.

The chart reveals distinct trends in the distribution of three paper types (C, J, and M) over the years 1990–2024. Paper type C experienced a steady rise in counts from 1990 to 2005, followed by a sharp decline starting in 2006, with no occurrences after 2021. In contrast, paper type J emerged in 2006 and demonstrated consistent growth, peaking in 2020, and maintaining high counts through 2024. Paper type M showed sporadic occurrences, with notable spikes in 2004 and 2007, but its presence diminished after 2015, with no counts recorded beyond 2020. The data suggests a shift in preference or production from paper type C to J over time, while paper type M remained relatively niche and inconsistent.

The chart reveals distinct trends in citation counts from Aminer and CrossRef over time. Aminer citations show a significant rise from the mid-1990s, peaking around 2006-2007, followed by a gradual decline, with near-zero citations by 2024. CrossRef citations exhibit a more moderate growth pattern, with a notable surge around 2006-2007, aligning with Aminer's peak, but they decline more steadily afterward. Both datasets highlight a sharp increase in citations during the early 2000s, suggesting a period of heightened academic activity or data availability. The decline in recent years for both sources may indicate reduced indexing, changes in citation practices, or incomplete data for the most recent years.

Building on the citation trends over time, the distribution of citations by paper type reveals that journal papers ('J') consistently receive the highest citation counts across both Aminer and CrossRef databases, with Aminer citations (125,403) significantly outpacing CrossRef citations (83,759). Conference papers ('C') follow, showing a notable drop in citations, with Aminer (144,530) still leading over CrossRef (36,335). Meanwhile, miscellaneous papers ('M') receive minimal citations in both datasets, with Aminer at 2,472 and CrossRef at just 805. This suggests that journals dominate in academic impact, while conferences and miscellaneous papers have comparatively lower citation influence, with a stark disparity between the two citation sources.

Further examining the impact of conferences on citations, the chart reveals that the 'Vis' conference has the highest total citation impact, with 120,100 Aminer citations and 37,925 CrossRef citations, indicating its strong influence across both citation sources. 'InfoVis' follows closely in Aminer citations (102,736) but surpasses 'Vis' in CrossRef citations (50,758), suggesting its prominence in CrossRef's domain. 'VAST' and 'SciVis' have significantly lower citation counts, with 'SciVis' being the least impactful, highlighting a steep drop in influence compared to the top two conferences. This suggests a clear hierarchy in the academic impact of these conferences, with 'Vis' and 'InfoVis' dominating.

The chart reveals a significant disparity in average citation counts across different awards, with papers associated with the "TT" award achieving the highest average citation count (281.2), followed by "TT;BP" (190.6). These values are substantially higher than the remaining awards, such as "BP" (57.859), "HM" (44.229), "BA" (42.833), and "BCS" (17.8). This suggests that papers recognized by "TT" and "TT;BP" awards may have a greater impact or visibility within their fields, while those associated with other awards tend to receive fewer citations on average, indicating a potential difference in prestige or relevance.

The chart reveals a significant upward trend in yearly downloads from 1990 to 2010, with a particularly sharp increase starting in the mid-1990s, peaking at over 200,000 downloads in the early 2010s. However, after 2014, downloads exhibit a gradual decline, with a sharp drop-off after 2020, reaching their lowest levels by 2024. This pattern suggests a period of rapid growth and adoption followed by a plateau and eventual decline, potentially indicating market saturation, shifting user preferences, or the emergence of competing platforms or technologies. The dramatic decrease post-2020 may also reflect external disruptions or a fundamental change in the product's relevance.

The chart reveals a significant disparity in downloads between papers with different award statuses. Papers in the "Unknown" category dominate with over 2.77 million downloads, far surpassing all other categories combined. Among awarded papers, "HM" (Honorable Mention) leads with 222,109 downloads, followed by "TT" (Top Tier) and "BP" (Best Paper) with 149,617 and 127,161 downloads, respectively. Categories with dual awards, such as "TT;BP," and less prominent awards like "BA" (Best Application) and "BCS" (Best Case Study), show drastically lower download counts, with "BCS" having the least at just 897. This suggests that while awards may enhance visibility, the "Unknown" category likely includes a broader range of papers that attract widespread attention, possibly due to factors beyond award recognition.