YouTube as a Source of Information in Understanding Autonomous Vehicle Consumers: Natural Language Processing Study

Abstract

The automotive industry is currently experiencing a revolution with the advent and deployment of autonomous vehicles. Several countries are conducting large-scale testing of autonomous vehicles on private and even public roads. It is important to examine the attitudes and potential concerns of end users towards autonomous cars before mass deployment. To facilitate the transition to autonomous vehicles, the automotive industry produces many videos on its products and technologies. The largest video sharing website, YouTube.com, hosts many videos on autonomous vehicle technology. Content analysis and text mining of the comments related to the videos with large numbers of views can provide insight about potential end-user feedback. This study examines two questions: first, how do people view autonomous vehicles? Second, what polarities exist regarding (a) content and (b) automation level? The researchers found 107 videos on YouTube using a related keyword search and examined comments on the 15 most-viewed videos, which had a total of 60.9 million views and around 25,000 comments. The videos were manually clustered based on their content and automation level. This study used two natural language processing (NLP) tools to perform knowledge discovery from a bag of approximately seven million words. The key issues in the comment threads were mostly associated with efficiency, performance, trust, comfort, and safety. The perception of safety and risk increased in the textual contents when videos presented full automation level. Sentiment analysis shows mixed sentiments towards autonomous vehicle technologies, however, the positive sentiments were higher than the negative.

The determinants of acceptability and behavioural intention of automated vehicles – a review

“…Results showed that there is no consensus among the public on trust in AVs (Salonen & Haavisto, 2019), with more studies highlighting a lack of trust in AVs than a high level of trust in them. For studies that undertook qualitative analysis, the word trust was one of the words with the highest frequencies, which implies a lack of trust in AVs, especially for frequent drivers (Cho & Jung, 2018;Das et al, 2019). The results also showed that trust may vary depending on the reliability.…”

“…The perceived safety varied according to the level of automation, but the meaning of this effect was interpreted differently. Some researchers showed that perceived safety decreased significantly with an increase in automation level (Das et al, 2019;Brell et al, 2019b;Nees, 2019) whereas other researchers found that perceived safety was higher for highly or fully AVs (Lee & Kolodge, 2019;Pettigrew et al, 2019). These results were corroborated by showing that AV technology was positively evaluated, created a perceived safety improvement and allowed drivers to be more relaxed (Lu et al, 2017;Kauer et al, 2015;Nishihori et al, 2017).…”

“…The second group corresponded to ethical, legal and safety questions. It included the following concerns: the safety of the software (Dong et al, 2019;Kyriakidis et al, 2015;Piao et al, 2016;Wicki & Bernauer, 2018;Zmud & Sener, 2017), the ethical questions raised by the introduction of the technology (Das et al, 2019;Hyde et al, 2017;Liljamo et al, 2018;Nordhoff et al, 2019), the need for new policies and legislation to regulate the use of AVs (Hyde et al, 2017;Kyriakidis et al, 2015;Lu et al, 2017;Piao et al, 2016;Wicki & Bernauer, 2018). Finally, the third group comprises the financial (Leicht et al, 2018;Nishihori et al, 2017;Piao et al, 2016;Ro & Ha, 2019;Shabanpour et al, 2018a;Shin et al, 2019;Wahlström, 2017;Zmud & Sener, 2017) and societal barriers, such as job losses (Nishihori et al, 2017;Nordhoff et al, 2019;Pettigrew et al, 2018aPettigrew et al, , 2019Wahlström, 2017;Wicki & Bernauer, 2018).…”

Case Study of Trend Mining in Transportation Research Record Articles

“…Another worthwhile study of topic modeling in the transportation field was conducted by Das et al (25) in which they studied topic changes of abstracts from papers presented at the TRB Annual Meeting from 2008 to 2014. Das et al used text mining and topic modeling in several other transportation studies (26)(27)(28)(29). Two recent studies used text mining and topic modeling TRB compendium papers and TRR papers (30,31).…”

Measuring the Effectiveness of Vehicle Inspection Regulations in Different States of the U.S.

“…In recent years, several studies have incorporated text mining in transportation engineering research: consumer complaint analysis ( 24 , 25 ), understanding public sentiments on safety enhancement and bike sharing ( 26 , 27 ), identifying research trends from transportation engineering conference papers and journals ( 28 – 32 ), crash narrative investigation ( 33 – 38 ), text analytics using social media mining ( 25 , 33 , 39 – 41 ), and transportation-related surveys ( 42 , 43 ). Ghazizadeh et al used text mining and latent Dirichlet analysis (LDA) to identify key patterns and non-trivial knowledge from the NHTSA complaint database.…”