Summary
The article “Visualizing Email Content: Portraying Relationships from Conversational Histories”, talks about a program called Themail. Themail is a visualization of the past emails you have. It grabs keywords from your emails and puts them into a visual portrait that you can use to visualize the kind of conversations you have with people. It has two layers, one of yearly words used frequently and another layer of monthly used words. They play on the size of words to help distinguish which words are used more frequently than others. Once they tested it out to the public, they were able to analyze that users either went with a haystack or needle method. The haystack method was more towards looking at things in a bigger picture. The users were more focused on the general kinds of conversations they were having with people. The needle method was more detail oriented and users wanted to find more specific conversations they had with people at specific times. At first people that followed the needle method were not impressed but when they were asked what kind of conversations they had, most users were able to remember and visualize what kind of conversation they had at that time. The application of this visualization is not for daily use because it uses an accumulation of emails from your past.

Reflection
This visualization is very interesting in how simple it is but how much information it gives to the users. It is simple parsing that grabs keywords that are used frequently from past emails; however, from this visualization, a user can get so much information of what kind of conversations and what kind of relationship they have with family and friends. It even allows two different uses. One that is more general and gives you a bigger picture of what kind of conversations you are having. The other view is more detailed and focuses on specific times and people you talk to. Both give you information on what conversations you have with that specific person and what kind of language you use. A simple visualization like this is amazing how little information is given but how meaningful that data is.

Questions
• Could this visualization be applied to other things than email like twitter?
• Is there a better way to represent the information in a different visual representation?
• If this is not really beneficial for daily use, is there a way we can make it useful to the user to keep checking?
• Is this information actionable or is it just interesting information people can get from using this visualization?

Read More

Summary 

This article discusses how email archives can be visualized to show individual relationships over a span of time based the frequency of and distinction of the words used in the interactions called Themail. They discuss other similar programs created to analyze email archives, but this is the only program that analyzes it based off the relationship between the user instead of just simply header information. Through their research they have found that the patterns displayed over time are significant to that persons life. This application is aimed at the end user instead of other sectors of the social network analysis community. Their visualization is comprised of two types of words: yearly and monthly. Yearly words are static and are large an grey, and faint in the background showing the most commonly used words over the year. Monthly words are interactive, yellow, and shown in the foreground as stacks on the timeline of the screen as a space. These words can be selected and interacted with to see the emails they showed in and other contexts. Additionally, there are circles that are exchanged in a month, the size of the circle indicates how lengthy the message was. The main functionalities this application tries to accomplish is to show the user (1) what do I talk about with a specific person in my email and (2) what are the differences in my conversations between people. This was used in user studies and case studies that gave positive feedback on this application, in which most users preferred the overall view (called haystack), instead of the detailed view you get from clicking on each word monthly (called needle). They liked the application to be able to see and remember their interactions with people. Some of the disadvantages to the way this application was created are that it does not differentiate between expressions, it only sees individuals words, limiting its accuracy in portraying an exact relationship in social terms we have learned as humans in society.

Reflection 

I understand why they chose to do the project and that there was positive feedback from users who volunteered to use the software. However, I do not actually see the value this applications brings to society. That might be because email is outdated, if they were to look at my vt email, it would just show the amount of work I have done, and not personal relationships, or travel records at the most. This applied to messenger applications might be interesting, but still I don’t know what value that brings, besides a emotional value. This emotional value of “looking back” is something a lot of companies deal with our online data have seemed to integrate into their applications. Both Facebook and Google Photos have a “rediscover this day” and “friend-aversairy”. Which at times I, as a user, enjoy when its applicable, sometimes they choose people I no longer have a strong connection without or funerals. I think the concept of reflecting on your life is interesting and something that is good for social media applications. I think this being applied to a different messenger might be something I could see more applications for, but honestly, now that the “rediscover this day” has been out a while, it has lost a lot of steam. I no longer share those anymore, I sometimes take a second to smile, or just simply think “oh yeah that happened”. But I don’t see a long term usage or need for an application like this. What value does this bring? Its the same concept that by the time Google Photos came out with “rediscover this day”, I was already over it. Because it doesn’t really add that much value, and multiple interfaces are reminding me of that event or lots of events. I don’t see a user sitting down at the end of every week/month and using this application to review what interactions they had because its not fulfilling an need they have. I see a user maybe every year, or every few years looking back at it, but then why develop this for an user to barely use unless its just for other purposes and other applications within the industry. This application extended and maybe integrated into the already existing thoughts of “rediscovering” a time, a relationship. or a place, might be of some use. I think users like it in the study like, cause I would like it and think its interesting, but not daily use applicable or even weekly or even monthly.

Read More

Summary

The article begins by describing email and the researchers’ past work with creating visualization for user’s email conversations. One particularly interesting part is how they mention that most people save all their emails, so this works great when doing research on the topic because there is plenty of data. The actual project they have created is called Themail, a visualization of a user’s email archive, they also performed research and found two prevalent themes among user reactions, the needle and the haystack.  Themail is set up as an interface that shows columns of words that are visually changed based on how much they are used as well as their distinctiveness in compared to other conversations, so bigger words are used often and uniquely in that particular conversation over time. Gaps between the columns indicate periods of time where no conversation occurred. The columns were also separated into two categories, yearly and monthly words.  Yearly words provided a broad overview of a relationship between two people over a large period of time, while monthly words indicate snapshots of events that occurred between those people. In the end it was found that the users they showed the application to, generally enjoyed using it and would like to have it integrated with their email client. Of the two modes, needle and haystack, many users liked to use the haystack mode. The haystack mode was the general use where you get to look at how your whole conversation with someone changed over time. A smaller number of people liked the needle mode, where you can look for specific words and find where they were said in that conversation. Overall, Themail was a successful experiment in the visualization of data.

Reflection

I think this was an interesting read because I like to read about how people go about implementing software. I also feel like it would be useful to have a piece of software that performs these functions. I especially would like to have the search feature, there have been a lot of times where I would need to find when a particular conversation happened. I don’t know how much use I would get out of the “haystack” portrait feature because most of my emails are for business or school and I don’t really have a lot of emotional emails to look back on, as that is what texting is for. The creators of the application definitely had a creative way to visualize the data though, as all the properties like word size and blank space are easily picked out from the visualization, and it is certainly clear what they represent. Visualization and user experience are fundamental when designing any kind of user application, this article represents that and also makes it more clear that our project needs to uphold these tenants in our design. However it is difficult to do this without user input, which is what the developers of Themail saw after users pointed out problems in the weight given to words in certain phrases.

Questions

• How could you go about visualizing people’s text message conversations? Would it be similar or significantly different?
• Why do you think more people enjoyed the “haystack” aspect of the program?
• Is there an alternative design for Themail that could be more effective?

Read More

Summary

The article “Visualizing Email Content: Portraying Relationships from Conversational Histories” is about Themail, a visualization that shows how people correspond to each other and how their relationship changes over time. It explains the parsing approach and interface, and further implications for research related to email content visualization. According to the article, users store a vast amount of emails ranging from insignificant to important. While many of these emails are insignificant, the pattern of emailing in which we develop becomes important. Two major themes are discussed: the appreciation of the overall picture, known as the haystack, and seeking specific pieces of information, known as the needle. Email visualizations fall into four main categories: thread based visualizations, social-network visualizations, temporal visualizations, and contact-based visualizations. Themail displays its information in yearly words and monthly words. Yearly words are used to show a broad overall tone of the relationship. This shows what people are usually conversing about, while ignoring words that are only used on special occasion. Monthly words are used to show a much more detailed portrait of past email exchanges. People were excited to look back at their email archives, and preferred looking at archives related to their family and loved ones.

Reflection

I believe that Themail is an extremely useful tool that should be used by all people that frequently email. Although I had never heard of it before, I would be very interested to interact with the tool to see the visualizations of my email archives. I think that this is something that should be more publicly broadcasted to the public, because many people do not delete their emails. People keep emails of conversations, news, and other relevant information that could be useful to analyze. It would be interesting to see this tool used on other platforms besides email. Many people have conversations on social media sites such as Facebook and Twitter, and rarely delete their messages. Seeing also how their results vary between platforms, analyzing whether or not they talk about certain topics more on specific platforms.

Questions

Are there any other visualizations out there like this?

Could this tool be incorporated into other social computing platforms?

 

Read More

Article: “Visualizing Email Content: Portraying Relationships from Conversational Histories”

Authors: Fernanda B. Viégas, Scott Golder, Judith Donath

Summary

The authors of this paper developed Themail a tool that parsed user email correspondences and displayed yearly and monthly keywords. They implemented two modes “haystack” and “needle”. Haystack mode displayed words on the screen that showed more general, big picture themes and trends. Needle mode picked out more minute detail. Most users preferred haystack mode, specifically reflecting on their discourses with family and loved ones. One of their methods included taking special measures to isolate and consolidate correspondences with specific individuals. That is, they made sure that if a user had several email addresses associated “John Smith”, the addresses were consolidated into a single contact.

Reflection

Themail is and interesting social tool. One might use it like a photo album or an evolving time capsule. It would be interesting to integrate other media and info such as photos, gifs, videos, location crossovers, routes, etc.

Questions

What would be the effect of “facebooking” this tool such that users were exposed only to correspondences that they were more positively engaged in?

Another leaf out of facebook – might users enjoy a video-graphic memoir of their meaningful correspondences?

Read More

Summary:

“The Language that Gets People to Give” is a paper written on what kinds of language and its use make for successful crowdfunded ideas. It begins with examples of both successful and unsuccessful ideas on Kickstarter, namely “Pebble” and “Ninja Baseball” and then attempts to figure out why one was successful, and the other not. Then it goes into depth on previous research in the field, and most importantly, why it is different or how it improves on the previous work. Then it goes over the different variables used and how the analysis was done. It then goes over the things that were found as the most important, pretty much all phrases, and shows a list of the phrases it found to be important. Finally it goes over other ideas that may help drive crowdfunding, such as giving out extra products as a reward for donating a certain amount, or giving out personal thanks.

Response:

This paper very quickly goes into detail about its analysis, and it’s actually quite understandable, especially compared to some of the stuff we’ve read before. This is good because language analysis can be difficult, and I know it’s something my group wants to do some work with. The fact that about half of projects get funded, where half obviously do not, is interesting. It’s just odd that it ends up being an almost perfect 50/50 split. The controlling for genre or category specific ideas and phrases is incredibly important to me. As a group working on figuring out false news articles it is integral that we be able to pull out things that aren’t necessarily part of the article, or phrases that are just used in the subject. Because of this, it’s something we’ll need to think about quite a bit. The statistics on adding the controls and phrases to the model is also incredible, 2.4% error is insanely low for almost any model. The table of phrases as well as the public data set may end up being very useful to our project as well. Even if they won’t be the same, because they won’t ours is dealing with news and informative articles not crowdfunding, the ideas are very similar. So the detail on the analysis should prove extremely useful. It’s also good to know that our teacher understands language analysis, I’ll be asking you more questions soon. Finally some of the alternative reasons for success could be useful to think about on the news standpoint. Each of them translates at least slightly, and if nothing else are things we should bear in mind and start thinking of things which pertain specifically to our subject (such as what sources an article cites, or what type of articles it cites).

 

Questions:

What other than language might determine how successful a crowdfunded idea is?

Have you ever participated in a crowdfunding? What was it for? Do you think it was a good idea to support it?

Read More

Summary:

In the paper “The Language that Gets People to Give: Phrases that Predict Success on Kickstarter,” the authors look into the power of phrases and words to encourage users to users to donate and use them as an indicator for the likelihood of a project being funded or not. They mainly focus on the language used in the project pitch but factor in other variables such as amount requested and timeline for the project. However even with the inclusion of other variables they found that the top 100 predictors were still phrases. The phrases were also classified into 13 overall categories: art, music, publishing, design, film and video, tech, dance, theater, photography, food, games, fashion, and comics. Each had respective sub-categories and the phrases were analyzed to see if they were exclusive to a single category since those would be useless and skew results. A lot of what the paper also found supported my experiences as well. Users love getting something in return and a lot of kickstarters will offer “limited editions” or “exclusive” items in return for reaching certain levels of donation. This is actually added to another grouping of categories: Reciprocity, scarcity, social proof, social identity, liking, and authority. All of which are widely used tactics in selling things to people

Reflections:

I will start by saying that throughout the paper I was wondering if I could take advantage of these phrases to try and get myself funded for something but then I realized that these are just indicators. I did take not of a lot of the analysis that went on in the paper, more so than previous ones. This one actually had very descriptive explanations of what each analysis was doing and how they worked. So not only was it informative from a research standpoint but I found it to be very useful from a student one as well. I’m not sure how much the words really do play a part though. I feel like personal interest is the biggest deciding factor and people are often willing to invest more in things that they like. So within six categories of persuasion was the information skewed towards one category over the others in statistically significant way? This article was still very interesting. We really do take everything for granted when the time and consideration people put into wordsmithing could easily be worth it.

Questions:

• Do these phrases hold true across all donation/crowdfunding sites?
• Do identical phrases using synonyms have the same impact?
• Inclusion of stretch goals and updates didn’t seem to be included but if they were to be would they have any appreciable impact?
• Do these phrases have any implications for advertising in real life? Could they affect in-person donation drives?
• Would this study be more useful if they were able to take into account all donations and how much each was for?

Read More

Summary:

The paper “The Language that Gets People to Give: Phrases that Predict Success on Kickstarter” examines the factors that contribute to the success of crowdfunding projects on Kickstarter. Kickstarter is a crowdfunding site for various projects to potentially earn all or no funding from online backers depending on whether the project’s funding goal is reached or not respectively. The paper builds off a similar prior study by looking at the language of the project pitch in addition to more project attributes. It was found that language such as word choice and phrasing played a large role in the success of crowdfunding projects. The phrases were selected by processing and filtering scraped project descriptions. Additional variables included project goal/duration, number of pledge levels, minimum pledge amount, featured in Kickstarter, video present/duration, etc. The phrases were categorized into six categories: reciprocity, scarcity, social proof, social identity, liking, and authority.

Reflection:

I found the results of the paper to be interesting and not too surprising since language is powerful and important if used correctly, especially since seemingly similar words have different conations. It is used all the time in marketing, so the results of the paper could be applied to other crowdsourcing site or products in general since the phrases were not selected for specifically Kickstarter. For example, the language aspect could possibly be applied to a product on Amazon or a channel on YouTube to increase success rates.

Questions:

• Are the phrases equally influential to all audiences/readers/consumers of the project?
• How different do the results vary depending on the site?
• Are the phrases only influential to the reader if the reader is not aware of the influence?
• How well does this apply to non-crowdsourcing sites?
• How many people read the project description completely versus just watching the video/pictures.
• Would first impressions or prior knowledge of the project affect the success?

Read More

Summary

The researchers of “The Language that Gets People to Give: Phrases that Predict Success on Kickstarter” explore what makes a Kickstarter campaign successful by looking at what phrases they use on their homepage. With their dependent variable being whether they were funded or not funded they also controlled for several variables such as having a video or how long the campaign lasted. They created a dictionary by scraping phrases used in a Kickstarter’s homepage using Beautiful Soup. They grouped the phrases they had scraped into meaningful categories using LIWC. They then ran a statistical analysis using penalized logistic regression on their findings. They found several trends that indicated success. For instance, if a Kickstarter offered reciprocity they were more likely to be funded. However, if they showed signs of doubt or used negative wording like “not been able to” they were less likely to be funded.

 

Reflection

Some of their results didn’t surprise me like their findings on reciprocity and social proof. A lot of it can be linked to psychology and sociology and how the behavior of others can affect how likely we are to participate in something. I think it would have been interesting to see, like the researchers suggest, how exploring the social network aspect can affect Kickstarter campaigns. It was interesting to see which phrases seemed to have a positive relationship with whether a campaign got funded though. There were some odd phrases like “dressed up” that weren’t obvious to me why that would make a negative impact. I also thought it funny that “cats” seemed to have a positive impact and could only reach the same conclusions the researchers did.

 

Questions

Would exploring more the reasoning of why the phrases have such an impact reveal any new information?

Would the same study on a different crowdfunding site reveal similar results or are these findings unique to Kickstarter?

Is there any correlation between the phrasing and the average size of donation?

 

 

Read More

Summary

The paper, “The Language that Gets People to Give: Phrases that Predict Success on Kickstarter”, discusses the types of phrases and the overall usage of language that can be found in successful crowdfunding project pitches. With a focus on Kickstarter, the authors conducted their research by scraping textual content from a collection of over 45,000 Kickstarter projects to gather over 20,000 common English phrases that were used in the project pitches and analyzed them with LIWC to categorize the phrases. From the data gathered, the authors found that the top 100 predictors of funded or not funded were all dependent on phrases. These predictors were found to exhibit various features of persuasion and were categorized into 6 groups: reciprocity, scarcity, social proof, social identity, liking, and authority.  At the end, all the predictive phrases found and the control variables were released a public dataset.

Reflection

This article was very interesting to read about as I never thought about how language could affect how successful a project could be. I have heard of crowdfunded projects and believed that the success of one was dependent on the idea that would be created but now I understand how the marketing of a product is just as important or more important than the actual idea. Furthermore, the correlation between certain phrases and factors of persuasion was eye opening. to learn how certain phrases could invoke factors of persuasion such as reciprocity and authority. As mentioned in the article, it would be interesting to examine how personal relevance could affect a project’s success. I believe the higher personal relevance, the more successful the project. Another possible direction for this project is to expand the research to other sites like Patreon, a subscription content service.

Questions

• Would the results be similar if this experiment was conducted using a different website?
• GoFundMe is another popular crowdfunding website but is focused more on getting funding for personal issues in comparison to Kickstarter, which is more geared towards professional projects. Would this difference have any effect on results of this research if it was performed on both sites?
• Were certain categories on the Kickstarter website more popular and successful than others?
• Are there cases where projects have been successful despite being an impractical idea due to well written pitches?

Read More