All posts tagged ‘sentiment analysis’

by CalebDecember 1, 2010

Littlecosm Games Twitter Sentiment

Littlecosm is a massively multiplayer Twitter client that is due out in January 2011. Being designed by Tokyo-based Yongfook, it promises to be an interesting mashup of real-time status updates, sentiment analysis, and dynamic gaming elements.

According to Yongfook's blog post:

Emotion plays a huge part in Littlecosm.  Littlecosm analyzes your tweets and figures out if you're in a good or bad mood.  If your tweets are mostly "bad", your character will turn out "bad" (you can see some "bad" characters in the screens above) - and vice versa.

Along with being a game powered by emotions, I wanted Littlecosm to be emotive....  I turned to nostalgia. Littlecosm is a game about collecting memories... Every memory you collect in Littlecosm has nostalgic value.  I think about how much fun I have talking to my friends about retro games, 80s music, tv shows from bygone eras, hypercolor tshirts... and the strange sense of discovery I feel when we talk about these things that we have already experienced.  That is the exact feeling I want to get from players of Littlecosm.

What Littlecosm looks to do is turn everyday behavioral data points into game mechanics. Yongfook is programming very deep human elements into this thing. Stamen's work for the VMA's and area/code's Sharkrunners are similar examples; both take dynamic real world data (people tweeting and sharks swimming) to game reality. Maybe it really is time to "program or be programmed".

Watch a quick demo of Littlecosm here:

by MBJanuary 29, 2010

iPad 1.0: Where and How We'll Use It


Who doesn't have something to say about the iPad? The sentiment seems to be split down the middle with some shrugging it off as an oversized iPod Touch and others hailing it as the future of computing.  (And some just make fun of the name like they did with the Nintendo Wii.) We took some time to analyze the overall sentiment on the web, as well as our own thoughts on the device.

iPad's Positioning


As Steve Jobs began to introduce Apple's newest product line, he described a space between the phone and the laptop where the iPad would rest comfortably (just like your knees). Many people were hoping for an all-in-one laptop replacement. Instead, the iPad is hoping to fill a space not yet satisfied, one that we still might not fully grasp. Because this space is new, many of us fail to imagine it exists, or even that our behavior might change as the iPad evolves.

While obviously lumped into Apple's "mobile offering," it will not replace our most important device -- the phone. Most of the buzzing technologies found in mobile will not have the same effect on the iPad. For example, mobile crowdsourcing will not be possible on the same scale, as the size will prevent us from taking it places our phones go. Same goes for location-based services. Without a camera, augmented reality can be scratched off the list. SMS is a no go and same with phone calls (although Skype is a possibility).

What we see is that Apple has clearly positioned the iPad between a Macbook and an iPhone by both design and marketing. They are creating a device to address needs that we are possibly unaware of, and create ones that still do not exist (Apple is in the business of selling products, after all...).

Where We'll Use It


Size matters. For the iPad, it will play a big part in where and how it will be used. As we mentioned earlier, the 9.7 inch screen means that we will not carry it everywhere we go, like we do a mobile phone. Instead expect to see it used  in environments where we have decided to settle for the time being -- the living room, kitchen, park bench, office, Starbucks, subway commute, classroom, airplane, etc. There was a reason Steve Jobs was lounging back in an armchair while presenting the device -- this is how Apple wants you to use it.

How We'll Use It


Keeping these contexts in mind, we can begin to imagine possible utilities or use cases for the iPad. Many of these include applications already on the iPhone that would make more sense with a larger screen i.e., reading, watching video, looking at pictures, surfing the web, and gaming. Looking further we see even more possibilities tied to contextually-affected behaviors. Many people use their laptop while sitting in front of the television; perhaps the iPad could perform as the ideal second screen for the 10-foot experience. Magazines and newspapers can create a more graphical, interactive digital experience than is possible on the Kindle or eReader. Cooks can bring it into the kitchen to read recipes online. Like a Chumby, the iPad could display ambient information and entertainment. It could reside on the coffee table and be used for two-screen interactions and real-time social viewing. Networks and TV should begin considering ancillary content that can be displayed on this device.

Mashable suggests that the iPad is made for consuming content and not creating it, we disagree. Just look at the creation tools for the iPhone -- piano apps, paint pad, remix tools... From Creative Applications:

Some people may be disappointed that it’s ‘basically a big iPhone’, I don’t think they realize the potential for a whole new breed of  multitouch applications, and a slew of new usage scenarios. The simple addition of a larger screen (and a faster processor) allows for much deeper applications that just weren’t possible on the iPhone.

It could also be used in the workplace and academic settings for note-taking, email, keep tabs on your calendar, and making presentations. Startups like Inkling have already popped up to provide for iPad-based content creation, powerful educator tools, and a richer learning experience.

Explaining Reaction and Disappointment


Many people are disappointed with what they saw at the unveiling, but British actor Steven Fry reminds us of the initial reaction to the iPhone.

In June 2007 when the iPhone was launched I collected a long list of “not impressed”, “meh”, “big deal”, “style over substance”, “it’s all hype”, “my HTC TyTN can do more”, “what a disappointment”, “majorly underwhelmed” and similar reactions.

He goes on to point to the three billion apps that have been downloaded in two years and labels the iPad as a John the Baptist, preparing for what is to come. Noah Brier's brings his thoughts on bloggers' disappointment:

The disappointment thing is pretty amazing. The specs are pretty much exactly what everyone expected (minus a camera maybe) and yet the response seems to be that this device isn't all that exciting (at least not yet). So what is it? Maybe we just don't get enough real mystery any more ...


Some don't see it in their daily routine, because maybe they really don't need it. Not everyone will. Keeping in mind that this is a first generation device and looking into the future, this is a good first step. Problems that will hopefully be addressed include the inability to multitask, the lack of Flash (which means no Hulu or Amazon video), no camera (there goes video conferencing), and of course the nagging problem of discovery in the App Store.

Like MG Siegler and Steve Jobs noted, "If you have an iPhone, you already know how to use this." Of course, this is just a platform and the real use cases will spring from the applications developed for it. But let's consider today's children who are being handed iPhones instead of pacifiers. They are growing up with a touchscreen interface -- no mouse and no keyboard. Will this be the computer of their future?

by CalebDecember 15, 2009

Humans as Sensors: US Geological Survey Measures Earthquakes With Twitter


The US Geological Survey has been tracking tweets to get instant public reaction to earthquakes. Similar to sentiment analysis and mobile crowdsourcing, the tactic involves scraping public opinion, and putting it to use in ways not originally intended. Tweets with geo-data attached help in more accurately pinpointing location.

Through harnessing the masses, the USGS is able to graph spikes in Twitter traffic and assess whether or not a situation is serious enough for deploying emergency responders. While the strategy is not entirely dependable (results for the videogame Quake clutter the search), it is yet another case for the power and flexibility of social media.

[via bbc]

by CalebNovember 5, 2009

TREND: Mobile Crowdsourcing - Using Humans as Sensors


Ever since the dawn of web, businesses, researchers, and even artists have been experimenting with crowdsourcing. Tapping the wisdom of the crowds is not just a smart PR and CRM strategy, it's a much more efficient way to collect information. Sentiment analysis, Aardvark's human search engine, and Victor & Spoils are all examples.

Mobile opens up a new angle on crowdsourcing, one that takes on a more stealth approach. While social media thrives on active participation, powerful data is being collected via our mobile devices while we, for the most part, remain unaware. Around the clock, our phones are collecting real-world information, mapping social processes, and transforming us into human sensors.

We naturally navigate the world using five key senses: sight, hearing, taste, touch, and smell. With these we construct the realm we live in. In the past, people have discussed ways to integrate these with computers, or build on them using technology. You could say that this is already happening. The current iPhone has about ten sensors including GPS, a compass, microphone, and camera. Smart phones can collect and process the data we are sensing 24/7, even measure the impact of ads we see and hear. With a guided purpose, businesses are using this to observe and innovate while we passively contribute information from our traveling selves.


We've covered Sense Network's Citysense, a startup that uses points of GPS and WiFi positioning data to view San Francisco's live activity and hotspots. After a quick download, data is collected automatically without need for further user action. With this it is possible to view what days of the week are busier, what venues are crowded, and where everybody is going from your current location. (While companies like Foursquare and Buzzd are beginning to provide these popularity contests as well, they are based on active check-ins.) Cartogrammar takes users' photos to paint Harvard Square based on dominant colors of geo-tagged Flickr photos. This passive crowdsourcing is much like sentiment analysis, which takes the pulse of web users by translating expressed emotion into hard data.

As with Citysense, mapping is natural application of this. Companies like Waze, Google, and Biketastic are crowdsourcing the actual creation of real-time traffic reports. Using previous speed and direction data, collected via mobile, Waze can predict future highway buildup and explain unusual situations.

This new forecasting and mapping is also applied to crowd flow through retail and other indoor locations by Sense Networks and Path Intelligence. The latter sets up receivers across a specified area that catch the frequencies sent out as mobile phones ping their towers. With this, they can observe movement through a building, saw a mall for example. This has powerful implications for everything from architecture and urban planning to advertising and marketing.

With data coming in every second of every day, our modern society now has a real-time pulse to be monitored. Like weather forecasters, employees at companies like Path Intelligence are able to use this to predict the future as well as act on alerts and unusual occurrences.

Of course, all of this might spook consumers just a bit. A recent report from Berkeley and UPenn showed that between 73 percent and 86 percent of adults opposes tailored advertising when they found out how it was done. Those techniques include tracking behavior on websites and in retail stores.

Considering this data is all anonymous and not tied to user profiles, that should quell concerns some. But expect to see much ado about this practice in coming years, as well as many more mind-blowing applications of it.

by MBAugust 25, 2009

Getting Sentimental: Mobile as Real-Time Emotional Gauge


Machines' ability to interpret human language continues to develop and is finally to a point where it can be put to use in powerful ways. There are automatic translators, simple spelling and grammar correctors, chat robots, complex search engines and now there is sentiment analysis, the process of turning human emotion into cold hard data. Using algorithms, computers can mine the mountains of online content and present a report of how the overall population feels about anything from a political campaign to that new burger on the menu in the Dallas test-market.

Companies see value in being able to provide this focus group 2.0 and are developing tools accordingly. Both Scout Labs and Jodange scan forums, social networks, blogs, and mainstream news to come up with opinion data. NBC's O&O sites asks visitors to vote on how they “feel” on all the story pages and then this is displayed via text that appears after pages load. Twitter has been used to analyze sentiment since Summize was experimenting with it in its labs back in 2008. Tweetfeel gives search results a sad or smiley face accordingly. Algorithms are intensely complex because of the linguistic and cultural factors including irony, sarcasm, and slang.

All of this is only made possible with the rise of social media where everyone and anyone can publicly comment and voice their personal opinion in a perfectly prepared package of digital characters. The ubiquity of mobile communication technology is supporting the real time web and allowing on the spot feedback. During Michael Jackson's funeral there were reportedly 5,000 Twitter updates per minute and on Facebook 6,000 status updates per minute.

Mobile technology is driving all of this microblogging and lifestreaming, which is essentially reams of sentiment-ladden data. Mobile is the collection mechanism--like an omnipresent therapist's couch, it invites confession. It does this by breaking down the barriers of inconvenience, time and place, the reasons people would originally avoid updating and sharing. A great example of mobile as real-time emotional gauge is the Track Your Happiness project out of Harvard. Through a mobile website and iPhone app, you can chronicle your emotions and see trends in the factors that affect them. As people continue to stream their lives digitally, companies will eventually learn to forecast, much like the Weather Channel, what reactions to certain happenings will be.

- Caleb Kramer