TechSized

Sense Networks is making sense of real world location data

jdb — Wed, 28 Jul 2010 06:34:50 +0000

Intro

We know that a user’s spatio-temporal data (i.e. location based + time) can provide good insights into a users behavior. However, as we start to go beyond a particular individual and try to tie similarities across users so as to develop user “types” using various sources of spatio-temporal data the task can become exceedingly complex. This is where Sense Networks step in. They have built out a solution based on a proprietary MVE algorithm that can take in these multi-dimensional data and reduce it using distributed (Hadoop) technology to develop connections that can meaningfully be extracted from such large amounts of data.

What they offer

Figure: describes the process of rationalizing data into specific profiles. Image is courtesy of Sense Network

Based on a presentation by Girish Rao, Product Manager from Sense Networks (at the NY Location Based Apps Meet-up), seemed like there were two key data domains that Sense Networks was focused on:

Location based data mining – Example – Cabsense NYC is an iPhone app that Sense Networks developed that identifies the best corner to stand on to catch a NYC cab. They derived this from the NYC’s Taxi and Limosine Commision’s (TLC) data from the Taxi’s GPS location (good thing most NYC cabs have GPS data).

User “type” identification – Based on data from the telcos (i.e. location based on tower, etc), Sense Networks has started to map out user profiles and as a by-product of that can help understand user preferences. Add to this financial data (if available), and then powerful profiles can be created which can clearly help segment the user populations. The advantages of this from an advertising perspective are obvious. Additionally, these profiles can also be used to make recommendations to the user. The possibilities are close to endless.

What is their approach

According to Girish Rao of RealSense Networks, his company wasn’t necessarily relying on user generated data about their preferences or locations (like foursquare.com). The key reason perhaps is what he alluded to, which was that they currently source most of their data from telecom carriers (they have user data by cell phone towers they have used, etc). Each carrier’s data can run into the millions of users’ data and this means Sense Networks, unlike foursquare.com does not have to wait around to build a meaningful population of users to really start mining user behavior. They already have this data and with more data from the telecom companies, they could literally get meaningful data of millions of users overnight!

Conclusions

Although Girish didn’t get too much into the pricing model, it seemed more like they were eager to make this data available to the developers to leverage and they were looking to get the developer community engaged. So if you are a developer looking to leverage this data, reach out to them http://www.sensenetworks.com/

No related posts.

Facebook – Its Social Graph API and its take on the Future of the Web

jdb — Sat, 16 Jan 2010 05:48:13 +0000

Facebook recently announced plans for the Social Graph API which allows external sites to host facebook features on their respective sites. This means instead of companies currently trying to establish a presence on facebook via facebook’s fan pages, external sites can now interact with users right on their respective web sites displaying other friends that may have visited their sites, left messages, etc. Liz Gannes of Gigaom.com calls this the “Dawn of Facebook’s people organized web.” She believes this will transform facebook into a hub that better connects people to the web (i.e. others, products, services, etc) via their own friends/friends of friends/larger circles. Although, this seems like a possibility, to what extent Facebook will revolutionize the way people get around in the web is still highly debatable.

Impacts

With the Social Graph API, external sites can now host facebook like features and publish and connect people to their social circles within the context of this external site. With the right amount of user participation one’s social circle may influence one to participate with the external site/offering, whereas previously that level of comfort to engage with an external site may not have existed. This is helpful for all involved parties. Additionally, retailers can track and offer incentives for people promoting their brand.

Hurdles

Google! While Facebook is starting to explore this new avenue of connecting people to the web via their friends (or friends or friends, etc), Google already has massive amounts of data about its users via its own services such as gmail, Google Search, Subscriptions, Android mobile. Google also houses massive amounts of data over the web and via its Ad Words (advertising), it’s own services, and the web it pretty much has more than enough information to connect a users’ or demographic’s pattern to better understand how best to serve the web to its users. See the illustration below.

Facebook's Vs. Google's Strategy

Participation is a critical component if facebook’s Social Graph API strategy needs to take off. Firstly, organizations need to be interested in connecting with facebook. Some industries such as fashion, music, etc will greatly benefit from this and so they may be the first adopters. However, it will be important to have a lot of these web sites participate. Additionally, it is also important for consumers to participate. Most people would rather be seen not promoting any one specific brand. If organizations can figure out a way to incent people (i.e. reward them without making them look like an Amway sales person), then consumer participation may not be so much of an issue.

Conclusion

Facebook is definitely transforming itself into a more pervasive entity on the web with the Social Graph API. This opening up of facebook, where external sites now can now connect with individuals on a more personal level surely has the potential to transform the way customers interact with products and services on the web. However, Facebook has a long road ahead of it especially with Google way ahead in this race.

Artwork/Thumbnail by Gauldo on Flickr

No related posts.

Google’s Real-time Search Prediction

jdb — Tue, 15 Dec 2009 08:15:52 +0000

Just a couple of weeks ago, I was thinking the unthinkable – is Google so big that it can’t see beyond itself. Although, Google undeniably has a culture of innovation, it simply hasn’t been able to produce game changing trends on the internet, well ever since it came up with it’s search engine. However, being as big as it is, it is simply enough for Google to invest in those ever evolving trends to make a significant impact on those trends. Case in point, Google’s Real-Time Search. Although the real-time web has been evolving over the last year or two, Google’s entry into this market could potentially be game changing. Last week Google unveiled its Real-time search capability. This means if I typed in “Climate Copenhagen” I could potentially get information on exactly what was happening with the summit at Copenhagen at any given moment. Ofcourse, I didn’t search that, but I did search “Christmas Shopping Deals”. I still had better success executing a standard search – see image below.

Real-time Search Results

Standard Search Results

The feeds currently point to Twitter, Craigslist or other real-time content sources and now mostly produce meaningless results (as seen from the searches above). However, if you make an assumption that this real-time content will become more relevant over time, you could see how potentially Google’s Real-time search could provide anyone and everyone all the information they ever needed as it was happening. In order for this to happen Google working with others such as Twitter, Foursquare, etc will need to figure out an effective way to weed out noise.

Artwork/Thumbnail by dullhunk on Flickr

Youtube Direct – A great platform to power-up citizen video journalism

Cloud Computing – Why you should consider it for your business

jdb — Wed, 02 Dec 2009 07:57:59 +0000

Cloud Computing can be defined as an abstracted technology infrastructure that consists of a number of computers all linked together to support multiple business applications in parallel and to efficiently manage the use of resources among those applications. Applications developed to run on the Cloud infrastructure inherit the following benefits:

Downtime virtually non-existent – Since the cloud infrastructure consists of many servers, even if some of the servers go down, the business application can continue to work on any one of the other servers. This also means that cloud computing can be an integral part of a company’s disaster recovery strategy.
Increased capacity to scale – Since resources can be acquired dynamically at run-time the capacity to scale becomes almost instantaneous provided all the solution components are on the cloud.
Infrastructure completely outsourced – Since most cloud environments are provided by third parties, there is no need to manage the infrastructure. The nominal monthly pay-per-use fee takes care of the infrastructure support .

Besides these obvious benefits, the one area where there is a great deal of uncertainty is around data security and privacy.

Overall cloud computing should be seriously considered for any application that is looking to leverage the advantages cited above.

Some big players offering cloud-computing services – Amazon.com, Google, Microsoft, and IBM.

No related posts.

Realtime Web – Interesting companies in this space

jdb — Tue, 24 Nov 2009 08:03:21 +0000

The Real-time web is where the next wave of innovation is happening on the web. Real-time web is simply a term used to describe the content created in real-time on the web. It’s mostly realized by users armed with smart phones and/or users with an easier means to generate content in real-time. A challenging aspect of this development is that it’s harder to filter and deliver context sensitive data to the end user since everything is happening in the real-time (Even Google has not yet been able to deliver contextual “real-time” data. I’m sure they are working on it). Here are a few ventures at center of this “Real-time” web wave:

Twitter of course is the poster child for the real-time web. Twitter asks its users a simply question of “what’s happening?” and expects the user to provide an answer in 140 characters or less. The mass adoption of this service, has led to interesting outcomes. For example, The Bombay shootings last year the Hudson airplane crash were first reported on Twitter. As the traditional news media is loosing their relevance, the news media outfits are looking to redefine themselves and most of them have a presence on Twitter.

Ustream.com is another interesting venture that helps users stream live video based content. Ustream.com has already gained the attention of the large TV networks who are now contemplating streaming their content live via ustream.com

SubPly is company that specializes in providing captions for video feeds in various languages. The captions are produced by real-life employees who transcribe/translate a recording or a live feed for a fee. This technology is amazingly useful for enterprises, that struggle to share their video content. Since the captions are available, enterprise searches can pick up and index the text, which means your video repository is now fully searchable.

Foursquare is another company innovating in the social media space, connecting friends with each other around a geographical location by announcing their current physical location. There have been other players in this space, but foursquare has added a gaming element which has greatly encouraged users to sign-up and participate. Currently, the gaming element simply awards points and badges to participation, but the gaming element can be tweaked any which way to make the interactions more exciting and entertaining.

Artwork/thumbnail from dpstyles’s photostream on Flickr