Stritar's chronolog

A case study in agile development: the algorithm for Ljubljana Realtime's event discovery

Mon, 22 Oct 2012 20:01:16 GMT

When we started working on Ljubljana Realtime, we decided to approach it in an agile way. Amongst others, we wanted to use a few interesting lean concepts such as rapid development, Minimum Viable Product and the Build - Measure - Learn iterations. Less than two months later, the results are in, and they are very pleasing. The MVP in the shape of an activity map was developed in a few weeks, only to show there is a lot of social noise which will somehow need to be taken under control. But that's currently low priority, since the first pivot is already taking place, slowly shifting the focus from the rich map application towards an event discovery algorithm and stream. That's where I see the most potential of Ljubljana Realtime, and in the last weeks, that's where the most work was done.

Launch

Ljubljana Realtime event discovery engine uses Foursquare trending venues and geo-tagged posts from Twitter, Instagram and Flickr to discover what's happening in real life. At least 6 people checked-in on Foursquare or two different people tweeting or posting photos in a single hour could mean something is going on. These events are posted to Twitter and Facebook, with links to the posts. A few versions of this algorithm were already deployed, each one solving new problems, resulting in a few micro Build - Measure - Learn cycles in a single month.

Iteration 1: Foursquare, no duplicates

The first version of the stream (bot) was a simple one, at that point it was meant to work as promotion for the map. The only thing it knew how to do was to wait a few hours until it posted the same thing again. I think Foursquare checkins are alive for three hours, so if a trending venue was still trending after that time, new people had to checkin and the venue was still buzzing.

Problem: Plain, no real added value.

Iteration 2: Adding activity from other sources

When we were trying to make some space on the crowded map, we started grouping posts from Twitter and Instagram by the nearest Foursquare venue, which meant having less boxes on the screen. This turned out to be quite a complex thing to do properly, but it was worth the effort. On only a few occasions, one venue would have multiple posts in a single hour, and in most cases, that meant something was happening there. This provided another very interesting potential for the activity stream. Actually, it made the stream bigger than the map could ever be.

(I love it when such things happen, when you are trying to solve a problem, and it turns out there is much more hidden behind the resolution.)

Groupping posts by a venue. Did Ljubljana Realtime just discover an athletic meeting taking place?

The next problem: Activity in some venues, specially generic ones such as "Ljubljana" would trigger the stream almost every day. Similarly, some large venues, such as supermarkets, would be trending too many times on Foursquare.

Iteration 3: Balancing the posts

The algorithm needed an update, which would lower the amount of times when a venue would be recognized as an event, either on Foursquare or on other channels. At first I though about an upgrade which would set the amount of people or tweets needed to trigger the "event discovered" action for a specific venue. This would enable us to reduce the importance of some venues, but it would also require manual work. Luckily, we came up with another brilliant idea: the more times a venue is trending, the harder it is for it to be trending again, at least for the next few days. Automatic balancing.

Venues with the most discovered events. Generic ones, besides massive places, such as train stations, cinemas, squares and shopping centers are too dominating.

The next problem: At this point, we have launched other test instances of Ljubljana Realtime (Maribor, Zagreb and Zurich), to see how the system behaves in other environments. Some cities are bigger, some are smaller, which means they produce different amount of activity. Besides, different services are used differently in different cultures.

Iteration 4: Supporting local instances

Foursquare is big in Croatia (Zagreb), but not so much in Switzerland (Zurich), which means Zagreb Realtime's stream had a lot of Foursquare trending posts, while Zurich's had a lot of "Increased activity on Twitter and Instagram" posts. It was obvious that local instances needed different algorithms. While having an option to set the amounts which would trigger the post on a specific venue would be too much to moderate, having the same logic on a specific region could work. And it does. Zagreb now needs more people checked-in on Foursquare, while Zurich needs more unique people tweeting or sharing photos.

Number of discovered events by type (Foursquare vs. Twitter + Instagram) on each day. Foursquare trending venues are dominating Zagreb, while social streams are dominating Zurich Realtime.

The next problem: The basic algorithm requires two different people to tweet/post from the same location in one hour. In case of Zurich, this amount was set to three, but it turns out this situation happens rarely, around 10 times fewer than with two people, or only two to three times a day. Obviously not enough.

Iteration 5: Improving the "increased activity" weight

You can only have a whole amount of people tweeting in the past hour. Two or three. In our case, we needed something in the range of 2 1/2. The modified solution adds the number of posts divided by ten to the number of users, which means that currently, at least two people making at least three posts in an hour will determine a trending event in Zurich. This is not a perfect solution from the event discovery view, but it does what urgently needed to be done: prevent having too many tweets in the stream.

The next problem: we currently have four Twitter accounts that tweet events for these four cities. Our target was for each of them to make around 10 - 15 tweets a day, which seems like a number that is not spam. But how can a person see which of these events is THE event?

Iteration 6: Going super venue level 2

The latest version of the algorithm now recognizes two levels of events. An event (mostly 6 people on Foursquare, mostly 2 different people tweeting), and an outstanding event (around 12 people on Foursquare, around 4 people tweeting). Our goal was to make this super event happen only once a few days, on rare occasions two times per day, and it has already happened a few times.

Sometimes super events happen, with tens of posts in a single hour, such as the one for Philips Fashion week. These events definitely require more exposure.

The next iterations

At this point, I'm very satisfied with how the algorithm works, even though a few other modifications need to be done (specially to support different days of week specifics and behavior). By measuring what is happening, learning from that information and building the next releases based on that knowledge, the activity stream logic has come a long way from the initial version. Measuring is crucial, and rarely we have went to such extent to enable this in the widest way possible (e.g. the update to balancing the posts based on the previous events would be trivial by itself, but we wanted to log things that would happen but didn't happen, besides things that actually happened).

These cycles of Build - Measure - Learn can be a lot hard work, but they provide great results, which are also very fun and rewarding. Some people simply need to see how deep the rabbit hole is. Do you have any other interesting cases or experience with this approach?

Discover what's happening in Ljubljana in real-time

Fri, 28 Sep 2012 13:37:29 GMT

I've always been a big fan of the power of the crowds. When a mass of people can achieve much more than a few skilled individuals can. And ever since we've started playing with Twitter's API, I've been think about the possibilities of this magnificent data source. Besides Twenity, we've done a few other Twitter mashups like Twitter walls, but this wasn't enough. We wanted something more - geolocation. Displaying information on a map in real-time. But since there aren't that many tweets equipped with GPS coordinates, we needed to include other services for more diversity as well. Which we did, and Ljubljana Realtime, a social event discovery application, was born.

Launch

Data and services

Ljubljana Realtime currently feeds on four different services: Twitter, Foursquare, Instagram and Flickr. It would be great if we could add other services as well, but Facebook doesn't allow public geo search, Google+ doesn't support geo search at all, and other services either aren't appropriate or don't offer an API which would allow us to get their data.

These four services are visited once a minute, and all posts in a radius of around 5km from Ljubljana city center are found: tweets, Foursquare trending venues, pictures from Instagram and Flickr. A variety of information created with different purposes on different occasions.

The area covered by Ljubljana Realtime. Different services require different searches, based on maximum allowed radius.

The application

These posts are displayed on a map, which was the original idea for the prototype. The last hours of posts on Google Maps, which can be zoomed and filtered at will. But the whole display felt a bit chaotic (still does), since there are many posts in vicinity of one another. That's why we knew we need to group similar posts, and we did this by the post's nearest Foursquare venue. Then a funny thing happened: this simple solution enabled something magnificent, something that could be much bigger than the whole posts-on-a-map application.

Anyways, since geo location is heavily connected with mobile devices, the application is responsive and fully compatible with most smartphones. Perhaps there will be native apps as well at one point.

The stream

Back to the magnificent. Originally, a Twitter bot was intended to come with the application (I love making those), which would tweet all trending foursquare venues to promote the application. But this seemed a bit lame, we needed to add something cooler. Something that would add more value and detect an event before a 4sq trending venue would happen. This is where the mentioned grouping of posts by venue came in handy, and the logic is as follows: if two or more people publish form the same venue in a single hour, this could very well mean something's happening there. And in most occasions, this turned out to be true. Read further for more details.

Discovering an event before a trending venue on Foursquare.

The event discovery stream is available on Twitter and Facebook.

The problems

Besides unstable APIs, the biggest problem we are currently facing is the geolocation itself. GPS chips in mobile phones are often not accurate enough, so people are located tens or hundreds of meters from their actual location. Combine that with the amount of Foursquare venues out there (imagine tall buildings), and you can understand Ljubljana Realtime sometimes has problems with connecting a post to a venue. Not to mention duplicated venues. We've eliminated some of this effect by only using venues with a certain amount of checkins and different users, but this will surely be the greatest challenge the project is facing in the future.

An event which is not.

The results

Problems aside, in most cases, Ljubljana Realtime event discovery works great. In a week or so since it's online, it discovered many events that were happening in Ljubljana (night run to the Castle, an athletic meeting, one of the first iPhones 5 in Slovenia, a public garage sale in park Tabor, etc.), and on many occasions, it discovered these events before a trending venue happened on Foursquare. Which is great. The potential is obviously there, and newer, improved versions and algorithms will surely behave even better.

A few of the great discoveries Ljubljana Realtime made.

The plans

This project is being developed in an agile way, where the application's behavior is constantly being monitored and changes deployed rapidly according to discovered strengths and weaknesses. The MVP (minimum viable product) is there, and with a few minor modifications, Ljubljana Realtime will soon be ready to expand to other cities and regions. Now it's up to you to help us, and it's pretty simple. When something magical is happening on and you are tweeting about it anyways, be a sport and click the arrow to include your geolocation in the tweet. By doing this, you will help others to discover what's going on in our beautiful city.

That's it for now, party on.

p.s. for all the Slovenians out there: the coordinates embedded in a tweet are pretty accurate even though Twitter will say you are in Italy. If you look at the picture of the map below the tweet, there's a polygon around Italy which sadly contains Slovenia as well. Hopefully, Twitter will remove bug someday.

The Silicon Valley tour, part 8: Lessons learned, time to reevaluate

Sun, 22 Apr 2012 18:01:36 GMT

It's been about month since I've returned from Silicon Valley, so I've had plenty of time to think about what happened there. This time I went out of curiosity, hoping to get the idea of how things work in the global center of technology. The next time I will be there for real business, approaching the situation more systematically. San Francisco and Silicon Valley are a great place to visit for profiles such as myself, so there surely will be a next time, when a wiser version of me will be able to do some serious shit. And I'll be wiser also because I've learned my lessons this time.

Ideas are worthless, but execution also isn't all it takes

When I started being an entrepreneur, I thought The idea was everything. While innovation may be important, it's mostly worthless by itself. No matter how good or revolutionary your idea is, hundreds of people probably have a similar idea at the same time. As Oren Michels puts it: "Stealth mode is stupid, execution is what matters.". Ideas are worth nothing if they're not executed properly. But wait, we've done a pretty good job with Twenity, but that still wasn't enough to make it abroad. It seems execution isn't enough as well.

At this point I think the hardest thing to do is the next step, putting the product on the market, making it recognized. At least that's where we are now, that's the obstacle we are currently facing. But I can already predict that the next hardest thing would be to find the investors, and the next to scale the business. Idea > Execution > Marketing > Financing > Scaling is a complex lifecycle, and you have to control every element before you can say you've made it. That's why you should try (and fail) as many times as possible. With every new project you do, it will be easier to dominate the prior steps, allowing you to focus more energy on the next ones.

Competition is fierce, but the market is big

The (Seedcamp) startups I've met on my trip really do some seriously crazy shit. Some of the things were so awesome I've almost lost faith in being able to do something similar. But I managed to somehow get back in the game. Neolab may not be cutting edge (yet) on the emerging fields such as mobile, HTML 5 or responsive design, but that's really not that big of an issue, since we make things that do what they're supposed to. That's the Execution phase, but luckily, the Market has the capacity to absorb many things, if approached properly. Face it - there will always be companies who do a better job than you, but it's the Market who'll decide who makes it or not.

I attended a Salesforce conference for a few hours (thanks Davorin!), which attracted thousands of visitors. In case you don't know them, they are the leading provider of Software as a Service CRM solutions, offering a service and a platform for other developers. Their product is really awesome, but that doesn't mean similar products can't find their own market. Researching on this issue, I found a review of the 10 best known SaaS CRM solutions, all of them obviously able to survive, even though Salesforce dominates this segment. And I'm sure there are hundreds, if not thousands more that are able to coexist at the same time on the huge global markets. Which brings me to my next point:

Highlights from a massive Salesforce conference

User experience is the new competitive advantage

You need to find your Unique selling proposition and competitive advantage. Most of the above mentioned Seedcamp startups found it in superior user experience. Take something that works, and make it work even better. Today's users are quite willing to migrate to a new product / service if it turns out to be better than the current one. A great example of this philosophy is vox.io, which offers a similar service than Skype (online calls and chat), but simply does it better. And by better I mean using new innovative approaches to make something more simple and intuitive to use. The user experience segment will only become more important and challenging in the future, since new ways of interactions with software are emerging (touch, voice, thought?).

vox.io's chat with embeddable content proves chat can be reinvented

Selling and pitching is ok

When you mention Sales, many people get the impression of a door-to-door salesperson trying to sell you books. Or company phones ringing with people who make it hard to say no to. These things give sales a negative connotation. But it seems that's a cultural thing. In San Francisco, everybody is trying to sell you stuff, everybody's pitching all the time. When you ask somebody "What do you do?", the person is already showing you their product on his/her iPhone. Maybe not for the actual sell, but just to get some decent feedback. You shouldn't feel bad about trying to sell something, that how the system works. Just make sure that something is worth selling.

One of the parties at vox.io's. Everybody was pitching.

You're nothing without the contacts

I admit I was a bit naive this time. I hoped the fact me being a blogger and a representative of two startups would open a few doors, but this turned out to be wishful thinking. It's true I did manage to find a few contacts on the spot, which enabled me to do great things (such as visiting the technology bluechips and the Internet archive), but my way to the influential blogs was closed. I heard that they receive a few hundred submissions a day, so you need to make sure your pitch is awesome and you have backdoors to deploy it. Do your homework, the next time I take on a similar adventure, I'll try to schedule as many meeting as possible in advance. I've finally found a reason to pimp up my LinkedIn profile, and I was lucky enough that:

The Slovenian (startup) scene is alive and kicking

One of the most fascinating things that happened to me was the chance to meet the Slovenian diaspora in Bay Area, and the startup scene around it. Thanks to a few individuals, such as Jure, Tomaž and Andraž, a lot of things are happening around Stanford University and vox.io's San Francisco headquarters, and everybody is very welcoming to the new generation of Slovenian immigrants and other startups, trying to make it abroad. Thank you for your kindness, I feel Slovenia!

The Slovenian diaspora meeting at the Stanford University

Foursquare is the ultimate travel guide, so get a local data plan

Forget maps, Foursquare has become the ultimate travel guide if you're looking for companies, tourist attractions or happening. That's why you'll need a local data plan, which costs a few bucks a day. Pretty much everything is on Foursquare (besides, physical addresses are often very hard to find online), and trending venues can point you to the events worth visiting. I've never used Foursquare in such a way, and the critical mass of users reached around here surely provides massive value beyond the game.

What now?

That pretty much sums it up. This trip was a great experience, but it's time to get back to the drawing board. Rethink my strategies. Redesign Twenity's user experience. Reinvent Neolab's business model. Stay in touch with the people I've met. Find new contacts for the next visit. Adopt new technologies. Research new markets. Think global.

Well, at least I won't get bored.

Check out the complete The Silicon Valley tour series.

The Silicon Valley tour, part 7: A few exciting new business models that actually work

Sun, 25 Mar 2012 17:28:22 GMT

Visiting Silicon Valley enabled me too peek into the future a bit. Finding out about new technology trends, meeting disruptive new software startups and seeing fascinating new business models that are proven to work. There are even more futuristic products / services than the ones mentioned in this post, but these are the ones that made the biggest impression on me. And they all share similar competitive advantages: using smartphones and other mobile devices, canceling middlemen and supporting cashless commerce.

Mobile payments and Square

While Square may already be mainstream in the States, something like it is far away from being used in Europe, not to mention Slovenia. Square is focused on mobile payments, and is currently offering three different services:

a credit card reader that you can plug into your mobile device
an iPad register app for merchants, supporting invoices, payments and inventory
an app for paying with your smartphone

Square is on fire, and there are tons of places in San Francisco where you can use it. Electronic payment industry leaders such as PayPal are already looking for ways to conquer the same market.

A place that accepts all major credit cards and Square

Personal chauffeur and Uber

Uber can only be described as taxi on steroids, bypassing the dispatcher. Ok, it's actually a chauffeur. Using a mobile app, you make a request, and one of the drivers answers it. After that, you can see where the driver is, when he will pick you up, how your driver looks like, and get a direct link to call them. A guy in a suit driving a Lincoln picks you up, takes you wherever you want, and after it, you just exit, without paying. You get an invoice on your email, together with the map of the route, and the fare gets deducted from your credit card. It's a bit more expensive than a cab, but still, a pretty awesome service, that will be expanding to other countries soon.

With Uber, you get picked up by a Lincoln like this

Carsharing and Zipcar

The next generation of rent-a-car is called carsharing, and it's offered by Zipcar. No agents, just your card that you use to unlock the car you've rented. Everything happens online, where you make a request for a car, get directed to a designated pick-up area, and the car is already waiting for you there. Petrol is included in the price, and if the fuel tank drops below 1/4, you just use their credit card to top it up. Results: rent-a-car has never been easier, cheaper and less stressful.

One of the Zipcar pick-up places. They are everywhere.

bonus: Co-working and NextSpace

Co-working represents shared offices that people from different startups and companies can use. Besides being cheaper than renting your own office, you also get a chance to spend time with other people, share knowledge and ideas, which is something I found out to be invaluable. NextSpace is an established co-working space in San Francisco, with more than 175 members and packages ranging from a few hundred dollars monthly for a desk to around a thousand for a complete office. Something similar is already happening in Slovenia, so if you are interested, please visit coworking.si.

Tables, offices, conference rooms. All available with NextSpace.

Conclusion

Some of these business models might not work as well outside the curious and tech-savvy culture of Silicon Valley, but they can still help as a demonstrator of things that can be done and an announcer of things to come. Let's face it, in technology, Silicon Valley is still miles ahead, so if you're looking for trends, the smartest thing to do is just to go along with it.

Trademarks and logos are the property of their respective owners.

Check out the complete The Silicon Valley tour series.

Web 2.5: Looking for the missing link between Web 2.0 and Web 3.0

Wed, 10 Nov 2010 17:39:56 GMT

The great revolution of Web 2.0 is still here and well – it looks like it's not going anywhere. Actually, its name did change, making "social media" more widely used today, but I'm still sticking to the original Tim O'Reilly's term. It's been more than five years since that happened, and a few years ago Web 3.0 was also starting to get mentioned. Web 3.0 mainly stands for the semantic web, using mathematical algorithms and meta data for trying to understand the meaning of content. But the whole thing is getting kinda old, because we still didn't see any real great results or services online – or perhaps they just did not make it to the mainstream. So while we wait for that to unveil, we could discuss something in between.

Client: Smartphone

In my opinion, there are a few quite significant changes in the Web we're used to using today, compared to the widely spread classic social Web 2.0 that already became fully mainstream. And the thing that's mostly responsible for this transition, surely has to be the smartphone. The wide spread of smartphones brought us new kinds of services, solutions, interactions and user experience, all powered by a powerful hand-held device supporting web connectivity. And as I will try to explain below, these services go way beyond Web 2.0, which is traditionally still powered by a personal computer. If you ask me, the mobile web is not just a new way to access the World Wide Web, it might as well be the new generation of it.

New capabilities

Modern smartphones come equipped with a wide arrange of additional hardware capabilities, such as high-resolution cameras, various sensors, compasses, gyroscopes and probably the most important – GPS. This fact created an array of new online services, where the ones worth mentioning would definitely be location (geo) based networking and augmented reality. Most of us can probably still remember when first phones with their 300x200 pixel cameras came out, and look at what happened now. Not to mention the giant tablet army that's heading our way.

Real-time web

Real-time web is the driving force behind Twitter, a thing so cool Google adopted it too, while Facebook is having problems to get its information out of privacy. It represents giving information when it happens, not hours (days) later, like the traditional (online) media does. While real-time web might be one of the biggest favorites for the next generation of Web beyond Web 2.0, it's still pretty useless if real time is only on the publisher's side. Mobile devices bring real-time to consumer's side too, enabling real time interactions between millions of users, which creates even more publishers and information – more than a group of professional journalists is ever capable of creating, making this world truly global.

Location-based networking

Location-based networking enables socializing based on where you and your friends physically are. Original players, such as Gowalla and Foursquare (who aleady has more than 3 million users) are already being copied by the mainstream, such as Facebook, Twitter and Google (all three naming them Places). While these services are rapidly spreading, they also introduced another interesting thing into networking, the gaming component, which uses badges and achievements to create real-life role playing games.

Augmented reality

Combining a compass and a GPS allowed another new type of software, called augmented reality, mostly in the form of solutions that use the camera image and draw things on top of it. There are already a few interesting applications of it available, and the marketing potential of this concept is probably huge. Augmented reality also managed to turn the smartphone into a deadly gadget, and where governments probably spent billions researching similar weapons, a soldier equipped with a €500 smartphone and a proper piece of software could lead a team of warriors who can see each other through walls.

Mobile apps

The previous shift in IT was for corporate software solutions to go from desktop to the web (and beyond to the cloud). Funny the trend, as today software is going back from web to mobile "desktop" applications, called apps. This happened because mobile apps enable a much better user experience (they can access the above mentioned capabilities) and at the same time provide less data consumption (making them cheaper and faster). Besides being a billion dollar industry, apps might also create the next step in the evolution of (business oriented) software and IT.

Is the mobile web actually Web 2.5?

These are a few examples of features we haven't seen before, and I think they should be considered when thinking about significant generations of the World Wide Web. The transition from Web 1.0 to Web 2.0 was from static to dynamic, from publishing to sharing and interaction, and the next big step worth mentioning could really be the one presented.

I admit the title of this post is a bit misleading, since this new concept has almost nothing to do with semantic Web 3.0. But I hope my thoughts are not that way off, since I haven't noticed any real authority write about the mobile web as a next big step in the Web's evolution. Right or wrong, from my point of view as a software architect and developer, the mobile era brought us much more than just online access everywhere, it brought a new generation of software and the Web. Besides, we have to be aware that this trend of mobile domination will surely evolve even further and continue to grow in the future. And since the name Web 3.0 is already taken, it might as well be called Web 2.5. What do you think?