�ݺ�ߣshows by User: rschifan

�ݺ�ߣshows by User: rschifan / http://www.slideshare.net/images/logo.gif �ݺ�ߣshows by User: rschifan / Wed, 14 Dec 2016 22:35:30 GMT �ݺ�ߣShare feed for �ݺ�ߣshows by User: rschifan Mapping the Sensorial Layers of a City /slideshow/mapping-the-sensorial-layers-of-a-city-70152781/70152781 nyc-141216-161214223530
NYUs Center for Data Science Lunch Seminar, Dec 14th, 2016 Researchers have used large quantities of online data to study dynamics in novel ways. Consider the specific case of online networked individuals (e.g., users of Twitter, Instagram, Flickr). Can their social dynamics be used to build better tools for future cities? To answer this question, our research has focused on understanding how people psychologically experience cities. As a result, we have created new mapping tools that leverage senses and emotions, thus complementing the corporate idea of an efficient and predictable smart city, with the ultimate goal of reaching urban happiness. The work presented in this talk mixes data mining, urban informatics, and computational social science to show how a creative use of social media and network-generated data can capture the aesthetic, olfactory and sonic layers of our cities at scale.]]>
NYUs Center for Data Science Lunch Seminar, Dec 14th, 2016 Researchers have used large quantities of online data to study dynamics in novel ways. Consider the specific case of online networked individuals (e.g., users of Twitter, Instagram, Flickr). Can their social dynamics be used to build better tools for future cities? To answer this question, our research has focused on understanding how people psychologically experience cities. As a result, we have created new mapping tools that leverage senses and emotions, thus complementing the corporate idea of an efficient and predictable smart city, with the ultimate goal of reaching urban happiness. The work presented in this talk mixes data mining, urban informatics, and computational social science to show how a creative use of social media and network-generated data can capture the aesthetic, olfactory and sonic layers of our cities at scale.]]> Wed, 14 Dec 2016 22:35:30 GMT /slideshow/mapping-the-sensorial-layers-of-a-city-70152781/70152781 rschifan@slideshare.net(rschifan) Mapping the Sensorial Layers of a City rschifan NYUs Center for Data Science Lunch Seminar, Dec 14th, 2016 Researchers have used large quantities of online data to study dynamics in novel ways. Consider the specific case of online networked individuals (e.g., users of Twitter, Instagram, Flickr). Can their social dynamics be used to build better tools for future cities? To answer this question, our research has focused on understanding how people psychologically experience cities. As a result, we have created new mapping tools that leverage senses and emotions, thus complementing the corporate idea of an efficient and predictable smart city, with the ultimate goal of reaching urban happiness. The work presented in this talk mixes data mining, urban informatics, and computational social science to show how a creative use of social media and network-generated data can capture the aesthetic, olfactory and sonic layers of our cities at scale. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/nyc-141216-161214223530-thumbnail.jpg?width=120&height=120&fit=bounds" /> NYUs Center for Data Science Lunch Seminar, Dec 14th, 2016 Researchers have used large quantities of online data to study dynamics in novel ways. Consider the specific case of online networked individuals (e.g., users of Twitter, Instagram, Flickr). Can their social dynamics be used to build better tools for future cities? To answer this question, our research has focused on understanding how people psychologically experience cities. As a result, we have created new mapping tools that leverage senses and emotions, thus complementing the corporate idea of an efficient and predictable smart city, with the ultimate goal of reaching urban happiness. The work presented in this talk mixes data mining, urban informatics, and computational social science to show how a creative use of social media and network-generated data can capture the aesthetic, olfactory and sonic layers of our cities at scale.

Mapping the Sensorial Layers of a City from Rossano Schifanella

]]> 370 4 https://cdn.slidesharecdn.com/ss_thumbnails/nyc-141216-161214223530-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Detecting Sarcasm in Multimodal Social Platforms /slideshow/detecting-sarcasm-in-multimodal-social-platforms/67632985 sarcasm-acmmm16-161025150801
Sarcasm is a peculiar form of sentiment expression, where the surface sentiment differs from the implied sentiment. The detection of sarcasm in social media platforms has been applied in the past mainly to textual utterances where lexical indicators (such as interjections and intensifiers), linguistic markers, and contextual information (such as user profiles, or past conversations) were used to detect the sar- castic tone. However, modern social media platforms allow to create multimodal messages where audiovisual content is integrated with the text, making the analysis of a mode in isolation partial. We first study the relationship between the textual and visual aspects in multimodal posts from three major social media platforms, i.e., Instagram, Tumblr and Twitter, and we run a crowdsourcing task to quantify the extent to which images are perceived as necessary by human annotators. Moreover, we propose two different computational frameworks to detect sarcasm that integrate the textual and visual modalities. Results show the positive effect of combining modalities for the detection of sarcasm across platforms and methods.]]>
Sarcasm is a peculiar form of sentiment expression, where the surface sentiment differs from the implied sentiment. The detection of sarcasm in social media platforms has been applied in the past mainly to textual utterances where lexical indicators (such as interjections and intensifiers), linguistic markers, and contextual information (such as user profiles, or past conversations) were used to detect the sar- castic tone. However, modern social media platforms allow to create multimodal messages where audiovisual content is integrated with the text, making the analysis of a mode in isolation partial. We first study the relationship between the textual and visual aspects in multimodal posts from three major social media platforms, i.e., Instagram, Tumblr and Twitter, and we run a crowdsourcing task to quantify the extent to which images are perceived as necessary by human annotators. Moreover, we propose two different computational frameworks to detect sarcasm that integrate the textual and visual modalities. Results show the positive effect of combining modalities for the detection of sarcasm across platforms and methods.]]> Tue, 25 Oct 2016 15:08:01 GMT /slideshow/detecting-sarcasm-in-multimodal-social-platforms/67632985 rschifan@slideshare.net(rschifan) Detecting Sarcasm in Multimodal Social Platforms rschifan Sarcasm is a peculiar form of sentiment expression, where the surface sentiment differs from the implied sentiment. The detection of sarcasm in social media platforms has been applied in the past mainly to textual utterances where lexical indicators (such as interjections and intensifiers), linguistic markers, and contextual information (such as user profiles, or past conversations) were used to detect the sar- castic tone. However, modern social media platforms allow to create multimodal messages where audiovisual content is integrated with the text, making the analysis of a mode in isolation partial. We first study the relationship between the textual and visual aspects in multimodal posts from three major social media platforms, i.e., Instagram, Tumblr and Twitter, and we run a crowdsourcing task to quantify the extent to which images are perceived as necessary by human annotators. Moreover, we propose two different computational frameworks to detect sarcasm that integrate the textual and visual modalities. Results show the positive effect of combining modalities for the detection of sarcasm across platforms and methods. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/sarcasm-acmmm16-161025150801-thumbnail.jpg?width=120&height=120&fit=bounds" /> Sarcasm is a peculiar form of sentiment expression, where the surface sentiment differs from the implied sentiment. The detection of sarcasm in social media platforms has been applied in the past mainly to textual utterances where lexical indicators (such as interjections and intensifiers), linguistic markers, and contextual information (such as user profiles, or past conversations) were used to detect the sar- castic tone. However, modern social media platforms allow to create multimodal messages where audiovisual content is integrated with the text, making the analysis of a mode in isolation partial. We first study the relationship between the textual and visual aspects in multimodal posts from three major social media platforms, i.e., Instagram, Tumblr and Twitter, and we run a crowdsourcing task to quantify the extent to which images are perceived as necessary by human annotators. Moreover, we propose two different computational frameworks to detect sarcasm that integrate the textual and visual modalities. Results show the positive effect of combining modalities for the detection of sarcasm across platforms and methods.

Detecting Sarcasm in Multimodal Social Platforms from Rossano Schifanella

]]> 660 6 https://cdn.slidesharecdn.com/ss_thumbnails/sarcasm-acmmm16-161025150801-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Surfacing the Hidden Beauty of Flickr Pictures /slideshow/icwsm15-beautycrowdschifanella/48830017 icwsm15-beautycrowd-schifanella-150601081129-lva1-app6892
We propose to use a computer vision method to surface aesthetically pleasant pictures from the immense pool of near-zero-popularity items. [@ICWSM 2015, Oxford] [Transcript] Have a look at these two Flickr pictures. What’s the difference between them? They are both aesthetically pleasant photos, however their social exposure is very different. Picture A received a lot of attention as showed by the high number of favorites, comments, and views while picture B has very low social signals. These are examples of respectively popular and unpopular content. Considering the 100M creative commons Flickr dataset, we found that only 2% of the pictures have more than 5 favorites while the remaining 98% have five or less. The vast majority of the content lies below the surface of attention. While exploratory interfaces like the Flickr Explorer page tend to promote appealing and popular images, what happens to the unpopular content? Is it possible that all those 98M pictures do not deserve attention? In our work instead of focusing on the aspects that make an item popular we flip the perspective and we take a look at the quality of the unpopular content trying to surface the hidden gems of Flickr pictures. To infer image aesthetic quality we resort to crowdsourcing. We sampled 10K images within 4 topical categories (nature, urban, people and animals). Because we want to see how quality varies with popularity within each category we randomly sample popular, average popular and unpopular pictures. We design a crowdsourcing task where a user has to evaluate how beautiful is an image on a 5-grade scale. We ask at least 5 independent judgments per image. Even if there is a positive correlation between popularity and beauty (that means on average popular pictures are perceived as appealing) there are popular photos that have low aesthetic value. This is not surprising since popularity is not driven only by intrinsic quality as shown in the past. More importantly, several pictures with near-zero favorites have been judged as very appealing. Their relative amount is very low, making any random sampling strategy totally ineffective. Social signals like comments or tags are most of the time ineffective since unpopular items rarely receive social feedback that means you cannot use tags or comments to surface them. A possible solution is to look at the pixels. To surface the high quality content we use the crowdsourced dataset to learn a computational aesthetic framework based on compositional visual features. Such framework is able to automatically score images according to their aesthetic value. With this algorithm we ranked 9M unpopular Flickr photos. We then went back to the crowd and we asked the workers to judge the top 200 rated photos, the most beautiful according to our algorithm. Consistently across categories, the perceived beauty of the surfaced images is comparable to the most popular photos. ]]>
We propose to use a computer vision method to surface aesthetically pleasant pictures from the immense pool of near-zero-popularity items. [@ICWSM 2015, Oxford] [Transcript] Have a look at these two Flickr pictures. What’s the difference between them? They are both aesthetically pleasant photos, however their social exposure is very different. Picture A received a lot of attention as showed by the high number of favorites, comments, and views while picture B has very low social signals. These are examples of respectively popular and unpopular content. Considering the 100M creative commons Flickr dataset, we found that only 2% of the pictures have more than 5 favorites while the remaining 98% have five or less. The vast majority of the content lies below the surface of attention. While exploratory interfaces like the Flickr Explorer page tend to promote appealing and popular images, what happens to the unpopular content? Is it possible that all those 98M pictures do not deserve attention? In our work instead of focusing on the aspects that make an item popular we flip the perspective and we take a look at the quality of the unpopular content trying to surface the hidden gems of Flickr pictures. To infer image aesthetic quality we resort to crowdsourcing. We sampled 10K images within 4 topical categories (nature, urban, people and animals). Because we want to see how quality varies with popularity within each category we randomly sample popular, average popular and unpopular pictures. We design a crowdsourcing task where a user has to evaluate how beautiful is an image on a 5-grade scale. We ask at least 5 independent judgments per image. Even if there is a positive correlation between popularity and beauty (that means on average popular pictures are perceived as appealing) there are popular photos that have low aesthetic value. This is not surprising since popularity is not driven only by intrinsic quality as shown in the past. More importantly, several pictures with near-zero favorites have been judged as very appealing. Their relative amount is very low, making any random sampling strategy totally ineffective. Social signals like comments or tags are most of the time ineffective since unpopular items rarely receive social feedback that means you cannot use tags or comments to surface them. A possible solution is to look at the pixels. To surface the high quality content we use the crowdsourced dataset to learn a computational aesthetic framework based on compositional visual features. Such framework is able to automatically score images according to their aesthetic value. With this algorithm we ranked 9M unpopular Flickr photos. We then went back to the crowd and we asked the workers to judge the top 200 rated photos, the most beautiful according to our algorithm. Consistently across categories, the perceived beauty of the surfaced images is comparable to the most popular photos. ]]> Mon, 01 Jun 2015 08:11:29 GMT /slideshow/icwsm15-beautycrowdschifanella/48830017 rschifan@slideshare.net(rschifan) An Image is Worth More than a Thousand Favorites: Surfacing the Hidden Beauty of Flickr Pictures rschifan We propose to use a computer vision method to surface aesthetically pleasant pictures from the immense pool of near-zero-popularity items. [@ICWSM 2015, Oxford] [Transcript] Have a look at these two Flickr pictures. What’s the difference between them? They are both aesthetically pleasant photos, however their social exposure is very different. Picture A received a lot of attention as showed by the high number of favorites, comments, and views while picture B has very low social signals. These are examples of respectively popular and unpopular content. Considering the 100M creative commons Flickr dataset, we found that only 2% of the pictures have more than 5 favorites while the remaining 98% have five or less. The vast majority of the content lies below the surface of attention. While exploratory interfaces like the Flickr Explorer page tend to promote appealing and popular images, what happens to the unpopular content? Is it possible that all those 98M pictures do not deserve attention? In our work instead of focusing on the aspects that make an item popular we flip the perspective and we take a look at the quality of the unpopular content trying to surface the hidden gems of Flickr pictures. To infer image aesthetic quality we resort to crowdsourcing. We sampled 10K images within 4 topical categories (nature, urban, people and animals). Because we want to see how quality varies with popularity within each category we randomly sample popular, average popular and unpopular pictures. We design a crowdsourcing task where a user has to evaluate how beautiful is an image on a 5-grade scale. We ask at least 5 independent judgments per image. Even if there is a positive correlation between popularity and beauty (that means on average popular pictures are perceived as appealing) there are popular photos that have low aesthetic value. This is not surprising since popularity is not driven only by intrinsic quality as shown in the past. More importantly, several pictures with near-zero favorites have been judged as very appealing. Their relative amount is very low, making any random sampling strategy totally ineffective. Social signals like comments or tags are most of the time ineffective since unpopular items rarely receive social feedback that means you cannot use tags or comments to surface them. A possible solution is to look at the pixels. To surface the high quality content we use the crowdsourced dataset to learn a computational aesthetic framework based on compositional visual features. Such framework is able to automatically score images according to their aesthetic value. With this algorithm we ranked 9M unpopular Flickr photos. We then went back to the crowd and we asked the workers to judge the top 200 rated photos, the most beautiful according to our algorithm. Consistently across categories, the perceived beauty of the surfaced images is comparable to the most popular photos. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/icwsm15-beautycrowd-schifanella-150601081129-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds" /> We propose to use a computer vision method to surface aesthetically pleasant pictures from the immense pool of near-zero-popularity items. [@ICWSM 2015, Oxford] [Transcript] Have a look at these two Flickr pictures. What’s the difference between them? They are both aesthetically pleasant photos, however their social exposure is very different. Picture A received a lot of attention as showed by the high number of favorites, comments, and views while picture B has very low social signals. These are examples of respectively popular and unpopular content. Considering the 100M creative commons Flickr dataset, we found that only 2% of the pictures have more than 5 favorites while the remaining 98% have five or less. The vast majority of the content lies below the surface of attention. While exploratory interfaces like the Flickr Explorer page tend to promote appealing and popular images, what happens to the unpopular content? Is it possible that all those 98M pictures do not deserve attention? In our work instead of focusing on the aspects that make an item popular we flip the perspective and we take a look at the quality of the unpopular content trying to surface the hidden gems of Flickr pictures. To infer image aesthetic quality we resort to crowdsourcing. We sampled 10K images within 4 topical categories (nature, urban, people and animals). Because we want to see how quality varies with popularity within each category we randomly sample popular, average popular and unpopular pictures. We design a crowdsourcing task where a user has to evaluate how beautiful is an image on a 5-grade scale. We ask at least 5 independent judgments per image. Even if there is a positive correlation between popularity and beauty (that means on average popular pictures are perceived as appealing) there are popular photos that have low aesthetic value. This is not surprising since popularity is not driven only by intrinsic quality as shown in the past. More importantly, several pictures with near-zero favorites have been judged as very appealing. Their relative amount is very low, making any random sampling strategy totally ineffective. Social signals like comments or tags are most of the time ineffective since unpopular items rarely receive social feedback that means you cannot use tags or comments to surface them. A possible solution is to look at the pixels. To surface the high quality content we use the crowdsourced dataset to learn a computational aesthetic framework based on compositional visual features. Such framework is able to automatically score images according to their aesthetic value. With this algorithm we ranked 9M unpopular Flickr photos. We then went back to the crowd and we asked the workers to judge the top 200 rated photos, the most beautiful according to our algorithm. Consistently across categories, the perceived beauty of the surfaced images is comparable to the most popular photos.

An Image is Worth More than a Thousand Favorites: Surfacing the Hidden Beauty of Flickr Pictures from Rossano Schifanella

]]> 5343 29 https://cdn.slidesharecdn.com/ss_thumbnails/icwsm15-beautycrowd-schifanella-150601081129-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Smelly Maps: The Digital Life of Urban Smellscapes /slideshow/smelly-maps-the-digital-life-of-urban-smellscapes/48814321 icwsm15-smelly-maps-schifanella-150531190410-lva1-app6892
Smell has a huge influence over how we perceive places. Despite its importance, smell has been crucially overlooked by urban planners and scientists alike, not least because it is difficult to record and analyze at scale. Here we explore the possibility of using social media data to reliably map the smells of entire cities. [Presented @ICWSM 2015, Oxford] [Transcript] Think about your nose. Now think about big data: you probably didn’t realize it yet but your nose is a big data machine. Humans are able to discriminate more than one trillion different odors. On one hand, we have our big data nose; on the other hand, we have city officials and urban planners who deal only with the management of less than ten bad odors out of a trillion. Why this negative and oversimplified perspective? Smell is simply hard to measure! How do we capture odors? We teamed up with Kate. She does her PhD at the royal college of art and her work is about smell walking. She collected smell-related words with smell walks in seven cities in UK, Europe, and USA. Locals were asked to walk around their city, identify distinct odors, and take notes. Smell descriptors were taken verbatim from the smell walkers’ original hand-written notes. To create a bridge to the online world, we matched the smell related words from Kate’s work with social media data. To structure this large and apparently unrelated dataset of smell words, we built a co-occurrence network where nodes are smell words and undirected edges are weighted with the number of times the two words co-occur. Communities are found to be internally coherent and they can be grouped in semantically related smell categories. Surprisingly this categorization reflects both negative and positive smells. Our categorization is valid for two main reasons: first of all the emergent top-level categories match with the classification found in the work of Victoria Handshaw Second, our smell categories are ecologically valid. Cross correlations between categories are mostly zeros that means categories are orthogonal. Expected categories - nature & emissions - are correlated with each other. The result of this process is the FIRST urban smell dictionary containing 285 English terms. We capture the smell notes of the two cities using an analogy coming from the perfume industry. Think about a perfume, imagine spraying it on your wrist, the top notes are those perceived immediately, they are very intense and they evaporate quickly. After a while you will start to smell the base notes that are odors that stay for hours on your skin. Middle notes are something in between. City odors work exactly the same. Showing the fraction of Flickr tags that match each category, we saw that the base notes for Barcelona and London are emissions and nature. They give a high-level olfactory footprint of the city. We validated the results by correlating the presence of certain smell clusters and air quality indicators. ]]>
Smell has a huge influence over how we perceive places. Despite its importance, smell has been crucially overlooked by urban planners and scientists alike, not least because it is difficult to record and analyze at scale. Here we explore the possibility of using social media data to reliably map the smells of entire cities. [Presented @ICWSM 2015, Oxford] [Transcript] Think about your nose. Now think about big data: you probably didn’t realize it yet but your nose is a big data machine. Humans are able to discriminate more than one trillion different odors. On one hand, we have our big data nose; on the other hand, we have city officials and urban planners who deal only with the management of less than ten bad odors out of a trillion. Why this negative and oversimplified perspective? Smell is simply hard to measure! How do we capture odors? We teamed up with Kate. She does her PhD at the royal college of art and her work is about smell walking. She collected smell-related words with smell walks in seven cities in UK, Europe, and USA. Locals were asked to walk around their city, identify distinct odors, and take notes. Smell descriptors were taken verbatim from the smell walkers’ original hand-written notes. To create a bridge to the online world, we matched the smell related words from Kate’s work with social media data. To structure this large and apparently unrelated dataset of smell words, we built a co-occurrence network where nodes are smell words and undirected edges are weighted with the number of times the two words co-occur. Communities are found to be internally coherent and they can be grouped in semantically related smell categories. Surprisingly this categorization reflects both negative and positive smells. Our categorization is valid for two main reasons: first of all the emergent top-level categories match with the classification found in the work of Victoria Handshaw Second, our smell categories are ecologically valid. Cross correlations between categories are mostly zeros that means categories are orthogonal. Expected categories - nature & emissions - are correlated with each other. The result of this process is the FIRST urban smell dictionary containing 285 English terms. We capture the smell notes of the two cities using an analogy coming from the perfume industry. Think about a perfume, imagine spraying it on your wrist, the top notes are those perceived immediately, they are very intense and they evaporate quickly. After a while you will start to smell the base notes that are odors that stay for hours on your skin. Middle notes are something in between. City odors work exactly the same. Showing the fraction of Flickr tags that match each category, we saw that the base notes for Barcelona and London are emissions and nature. They give a high-level olfactory footprint of the city. We validated the results by correlating the presence of certain smell clusters and air quality indicators. ]]> Sun, 31 May 2015 19:04:10 GMT /slideshow/smelly-maps-the-digital-life-of-urban-smellscapes/48814321 rschifan@slideshare.net(rschifan) Smelly Maps: The Digital Life of Urban Smellscapes rschifan Smell has a huge influence over how we perceive places. Despite its importance, smell has been crucially overlooked by urban planners and scientists alike, not least because it is difficult to record and analyze at scale. Here we explore the possibility of using social media data to reliably map the smells of entire cities. [Presented @ICWSM 2015, Oxford] [Transcript] Think about your nose. Now think about big data: you probably didn’t realize it yet but your nose is a big data machine. Humans are able to discriminate more than one trillion different odors. On one hand, we have our big data nose; on the other hand, we have city officials and urban planners who deal only with the management of less than ten bad odors out of a trillion. Why this negative and oversimplified perspective? Smell is simply hard to measure! How do we capture odors? We teamed up with Kate. She does her PhD at the royal college of art and her work is about smell walking. She collected smell-related words with smell walks in seven cities in UK, Europe, and USA. Locals were asked to walk around their city, identify distinct odors, and take notes. Smell descriptors were taken verbatim from the smell walkers’ original hand-written notes. To create a bridge to the online world, we matched the smell related words from Kate’s work with social media data. To structure this large and apparently unrelated dataset of smell words, we built a co-occurrence network where nodes are smell words and undirected edges are weighted with the number of times the two words co-occur. Communities are found to be internally coherent and they can be grouped in semantically related smell categories. Surprisingly this categorization reflects both negative and positive smells. Our categorization is valid for two main reasons: first of all the emergent top-level categories match with the classification found in the work of Victoria Handshaw Second, our smell categories are ecologically valid. Cross correlations between categories are mostly zeros that means categories are orthogonal. Expected categories - nature & emissions - are correlated with each other. The result of this process is the FIRST urban smell dictionary containing 285 English terms. We capture the smell notes of the two cities using an analogy coming from the perfume industry. Think about a perfume, imagine spraying it on your wrist, the top notes are those perceived immediately, they are very intense and they evaporate quickly. After a while you will start to smell the base notes that are odors that stay for hours on your skin. Middle notes are something in between. City odors work exactly the same. Showing the fraction of Flickr tags that match each category, we saw that the base notes for Barcelona and London are emissions and nature. They give a high-level olfactory footprint of the city. We validated the results by correlating the presence of certain smell clusters and air quality indicators. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/icwsm15-smelly-maps-schifanella-150531190410-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds" /> Smell has a huge influence over how we perceive places. Despite its importance, smell has been crucially overlooked by urban planners and scientists alike, not least because it is difficult to record and analyze at scale. Here we explore the possibility of using social media data to reliably map the smells of entire cities. [Presented @ICWSM 2015, Oxford] [Transcript] Think about your nose. Now think about big data: you probably didn’t realize it yet but your nose is a big data machine. Humans are able to discriminate more than one trillion different odors. On one hand, we have our big data nose; on the other hand, we have city officials and urban planners who deal only with the management of less than ten bad odors out of a trillion. Why this negative and oversimplified perspective? Smell is simply hard to measure! How do we capture odors? We teamed up with Kate. She does her PhD at the royal college of art and her work is about smell walking. She collected smell-related words with smell walks in seven cities in UK, Europe, and USA. Locals were asked to walk around their city, identify distinct odors, and take notes. Smell descriptors were taken verbatim from the smell walkers’ original hand-written notes. To create a bridge to the online world, we matched the smell related words from Kate’s work with social media data. To structure this large and apparently unrelated dataset of smell words, we built a co-occurrence network where nodes are smell words and undirected edges are weighted with the number of times the two words co-occur. Communities are found to be internally coherent and they can be grouped in semantically related smell categories. Surprisingly this categorization reflects both negative and positive smells. Our categorization is valid for two main reasons: first of all the emergent top-level categories match with the classification found in the work of Victoria Handshaw Second, our smell categories are ecologically valid. Cross correlations between categories are mostly zeros that means categories are orthogonal. Expected categories - nature & emissions - are correlated with each other. The result of this process is the FIRST urban smell dictionary containing 285 English terms. We capture the smell notes of the two cities using an analogy coming from the perfume industry. Think about a perfume, imagine spraying it on your wrist, the top notes are those perceived immediately, they are very intense and they evaporate quickly. After a while you will start to smell the base notes that are odors that stay for hours on your skin. Middle notes are something in between. City odors work exactly the same. Showing the fraction of Flickr tags that match each category, we saw that the base notes for Barcelona and London are emissions and nature. They give a high-level olfactory footprint of the city. We validated the results by correlating the presence of certain smell clusters and air quality indicators.

Smelly Maps: The Digital Life of Urban Smellscapes from Rossano Schifanella

]]> 1402 2 https://cdn.slidesharecdn.com/ss_thumbnails/icwsm15-smelly-maps-schifanella-150531190410-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Shortest Urban Paths or Shortcuts to Happiness? /rschifan/shortest-urban-paths-or-shortcuts-to-happiness schifanella-chasm14-140624223947-phpapp02
When providing directions to a place, web and mobile mapping services are all able to suggest the shortest route. The goal of this work is to propose ways of automatically generating routes that are not only short but also emotionally pleasant. To quantify the extent to which urban locations are pleasant, we build a crowd-sourcing platform that shows two street scenes in London (out of hundreds), and a user votes on which one looks more beautiful, quiet, and happy. Given such crowd-sourced ground truth, we are able to compute a proxy for the beauty dimension from Flickr metadata associated with more than 3.7M pictures in London. We finally arrange locations into a graph upon which we learn popular and pleasant routes. We quantitatively validate the extent to which the recommended routes are not only short (they add just a few extra minutes to the shortest routes) but also popular and emotionally-pleasing: compared to the shortest routes, they are always perceived as more pleasant, with an increase of up to 30%. We then qualitatively evaluate the recommendations by conducting a user study involving as many as 20 participants who have not only rated the recommendations but also carefully motivated their choices.]]>
When providing directions to a place, web and mobile mapping services are all able to suggest the shortest route. The goal of this work is to propose ways of automatically generating routes that are not only short but also emotionally pleasant. To quantify the extent to which urban locations are pleasant, we build a crowd-sourcing platform that shows two street scenes in London (out of hundreds), and a user votes on which one looks more beautiful, quiet, and happy. Given such crowd-sourced ground truth, we are able to compute a proxy for the beauty dimension from Flickr metadata associated with more than 3.7M pictures in London. We finally arrange locations into a graph upon which we learn popular and pleasant routes. We quantitatively validate the extent to which the recommended routes are not only short (they add just a few extra minutes to the shortest routes) but also popular and emotionally-pleasing: compared to the shortest routes, they are always perceived as more pleasant, with an increase of up to 30%. We then qualitatively evaluate the recommendations by conducting a user study involving as many as 20 participants who have not only rated the recommendations but also carefully motivated their choices.]]> Tue, 24 Jun 2014 22:39:47 GMT /rschifan/shortest-urban-paths-or-shortcuts-to-happiness rschifan@slideshare.net(rschifan) Shortest Urban Paths or Shortcuts to Happiness? rschifan When providing directions to a place, web and mobile mapping services are all able to suggest the shortest route. The goal of this work is to propose ways of automatically generating routes that are not only short but also emotionally pleasant. To quantify the extent to which urban locations are pleasant, we build a crowd-sourcing platform that shows two street scenes in London (out of hundreds), and a user votes on which one looks more beautiful, quiet, and happy. Given such crowd-sourced ground truth, we are able to compute a proxy for the beauty dimension from Flickr metadata associated with more than 3.7M pictures in London. We finally arrange locations into a graph upon which we learn popular and pleasant routes. We quantitatively validate the extent to which the recommended routes are not only short (they add just a few extra minutes to the shortest routes) but also popular and emotionally-pleasing: compared to the shortest routes, they are always perceived as more pleasant, with an increase of up to 30%. We then qualitatively evaluate the recommendations by conducting a user study involving as many as 20 participants who have not only rated the recommendations but also carefully motivated their choices. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/schifanella-chasm14-140624223947-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> When providing directions to a place, web and mobile mapping services are all able to suggest the shortest route. The goal of this work is to propose ways of automatically generating routes that are not only short but also emotionally pleasant. To quantify the extent to which urban locations are pleasant, we build a crowd-sourcing platform that shows two street scenes in London (out of hundreds), and a user votes on which one looks more beautiful, quiet, and happy. Given such crowd-sourced ground truth, we are able to compute a proxy for the beauty dimension from Flickr metadata associated with more than 3.7M pictures in London. We finally arrange locations into a graph upon which we learn popular and pleasant routes. We quantitatively validate the extent to which the recommended routes are not only short (they add just a few extra minutes to the shortest routes) but also popular and emotionally-pleasing: compared to the shortest routes, they are always perceived as more pleasant, with an increase of up to 30%. We then qualitatively evaluate the recommendations by conducting a user study involving as many as 20 participants who have not only rated the recommendations but also carefully motivated their choices.

Shortest Urban Paths or Shortcuts to Happiness? from Rossano Schifanella

]]> 782 8 https://cdn.slidesharecdn.com/ss_thumbnails/schifanella-chasm14-140624223947-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 https://cdn.slidesharecdn.com/profile-photo-rschifan-48x48.jpg?cb=1530435198 http://www.di.unito.it/~schifane https://cdn.slidesharecdn.com/ss_thumbnails/nyc-141216-161214223530-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/mapping-the-sensorial-layers-of-a-city-70152781/70152781 Mapping the Sensorial ... https://cdn.slidesharecdn.com/ss_thumbnails/sarcasm-acmmm16-161025150801-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/detecting-sarcasm-in-multimodal-social-platforms/67632985 Detecting Sarcasm in M... https://cdn.slidesharecdn.com/ss_thumbnails/icwsm15-beautycrowd-schifanella-150601081129-lva1-app6892-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/icwsm15-beautycrowdschifanella/48830017 An Image is Worth More...