Here's the transcript of our conversation:
RVB: 00:00:26.434 Hello, everyone. My name is Rik, Rik Van Bruggen from Neo4j, and here I am, again, recording another Neo4j Graph Database Graphistania Podcast. And tonight I am joined by two guys from Philadelphia in the USA that have been doing some amazing work with Neo4j. And on top of that, they have the funniest company name, I think at least. It's called Untitled Folder. That's Jess and Jason. Jess and Jason, welcome to the podcast.Subscribing to the podcast is easy: just add the rss feed or add us in iTunes! Hope you'll enjoy it!
JM: 00:01:04.071 Hi. Thanks for having us on the podcast.
JC: 00:01:04.525 Hello.
RVB: 00:01:06.187 Yeah. Thank you for joining me. That's really kind. Thank you. So, Jess and Jason, I don't know who wants to go first, but you guys have been doing some really interesting stuff together with Neo4j for the Philly community. Maybe you can just introduce yourselves and explain what you've been doing there.
JM: 00:01:27.488 I'll start. I'm Jess Mason. I work with my business partner Jason and we started a company called Untitled Folder, where we help start-ups or people with ideas that want to build an application but don't know how to use technology. We get them from idea to MVP. I believe every good idea starts with an Untitled Folder, so we want to figure out what they want to put in theirs and help them build their MVP, to launch it and get it into the hands of the customers as soon as possible.
JM: 00:02:02.112 I met Jason a few years back at a hack-a-thon, where we were talking about some ideas we were trying to work on. And I told him about this crazy technology called Neo4j and he said that he'd be willing to work on it with me and learn how to use it to build some recommendation engines I was trying to build back them. And say, I learned about Neo about 2011 and I didn't know how to use it at all and I wasn't really a technologist at the time. And so meeting Jason was a natural fit for us to work together. And during the hack-a-thon, we decided to hang out again sometime later and we started a business together. And now we run the Philly GraphDB meetup group. So we met our friend Karin who works at our co-working space and she said to us one day, hey, if you guys want to start the podcast go ahead-- the meetup group, sorry, not the podcast.
RVB: 00:03:08.332 You have it well if you want, you know [laughter].
JM: 00:03:12.361 We've been running the Neo Philadelphia meetup group for the last two years now and working on different projects together and growing our company.
RVB: 00:03:23.623 Great. Really great. And what about you, Jason? Maybe you can introduce yourself as well?
JC: 00:03:27.922 Yeah, sure. I'm Jason Cox. I've been doing full-stack web apps for a good while now, so integrating with Neo4j was a really good fit for me since I hadn't really go into the recommendation engine style that new modern apps are moving into and I just feel that it decreases the barrier to entry, to get into AI recommendation assist your customers style of apps that is really important these days, so that was my entry Jess showed me and got me into it and I've found, so far, it works exactly as I would have hoped.
RVB: 00:04:21.364 That's great. And I know you guys have done a bunch of projects with Neo4j already, but the most intriguing one from my perspective, at least, is the one that you've been doing for the local city community, right? And the data journalism around Cypher Philly. Can you tell us a little bit more about that?
JM: 00:04:45.130 Sure. We started Cypher Philly about seven or eight months ago after a meetup we had done working on journalism and using Neo4j. And we came up with the idea sometime around when we were doing that meetup. We said, "Hey, we have this powerful technology called Neo4j and people are using it for stuff like Paradise and Panama Papers and they can expose all this unrelated data and link it together using relationships in with the [inaudible]. The City of Philadelphia has a great web portal called Open Data Philly and it has a ton of data about the city and things that you would want to know that would help improve the city or just self-reporting on what's going on. And we thought, what a great civic idea to try and link this data together and do some investigative journalism, do some recommendation engines for the city, to improve the city in some way, and see what other people want to-- how they want to get involved.
JM: 00:05:53.825 So we started doing some meetups around this idea, about taking Neo4j and cyphering, literally, the open data that's available to us. So we started that as an umbrella project and it's grown over the last seven, eight months into a group of people, about 40-plus people that have wanted to join on and help in some way. And they're activists, journalist, coders, creatives. And people just generally want to get involved in improving the city.
RVB: 00:06:29.436 That's so fantastic to hear, yeah.
JM: 00:06:31.609 Yeah. And we worked with Code for Philly, their brigade for Code for America, and they launched some civic campaigns to try and grow these types of groups and help accelerate these ideas. And so we went there and gave a pitch about what we're doing and gained a whole lot more traction and have been partnering with them for a lot of projects and journalism. It's just been continually growing, so we've been getting sponsorship for it now, we even host servers from some people that have helped us out. We have meetup places we can go and it's just been incredible the amount of outreach we've gotten from it.
RVB: 00:07:17.306 Real grassroots, right? Really great. Yeah. Fantastic. So maybe this is a good segue to kind of understand why is Neo4j and Graph Databases, in general, why is it such a good fit for the type that you guys do, both in the Cypher Philly environment and professionally? Why is it such a good fit for you guys?
JC: 00:07:42.433 Well, mostly because it's really good at finding connections in a cross-data perspective that really you don't get with SQL or NOSQL. You basically can use Neo4j like it's SQL, you get your data and data objects and give it fields and use it that way. But making the cross connections of data and in a really generic, whole data sense gives a lot of flexibility, so specifically for helping journalist sort through public data, it's really just a great way to find connections, where normally departments wouldn't talk or you wouldn't look for connections between these types of data, so we can use a public source like Open Data Philly, get everything we can in there, and then find really amazing new connections.
JC: 00:08:58.872 And then, yeah, most every startup these days is really needing to and wants to get into recommending information for their users. Without that, it doesn't assist you using the tool the same, so Neo4j-- we basically did a hack-a-thon where we learned how to do recommendations at GraphConnect last year. And we had a recommendation running in like, two, three hours flat. It was really easy to use; the tools were provided for us. There's a web consult to try things out before you connect these data calls in your app. It just makes it really easy to get it set up and testing out and make sure it's working the way you're expecting it.
JM: 00:09:58.247 And that helps us for our customers, too, that are looking to get off the ground quickly and try these new technologies and also be able to scale with them later, so it's not these throw-away ideas or other applications you might try to integrate that cost a lot more as you scale. So the more you use them, the more they cost. But Neo allows you to work with your own data and the tools are open source, so it's great. You don't have to pay per transaction or query, which is a huge benefit to people we want to work it.
RVB: 00:10:30.890 Yeah. Absolutely. So it sounds like both from a functional point of view and the recommendation system approach that you're using also, and also the flexibility that the data model gives you is a real driver for what you guys are doing. Am I summarising that okay?
JC: 00:10:49.142 Yeah. It lets you build a real-world prototype. So yeah, you can change your mind along the way a lot easier. Yeah.
RVB: 00:11:00.548 That's really, really cool. So what's next for you guys? When is the Untitled Folder getting titled [laughter]? What does the future hold?
JM: 00:11:12.919 Yeah. I mean, right now we're working on a grant that we've been doing for the last few months. And I think we're in the final stages of that. We're just waiting to hear back whether we've got it. So we want to take that with Cypher Philly in the new year if we get it. And that will help us grow that out a lot more. We'd like to create some products from that that we can work with different organizations that could use those type of services.
JC: 00:11:40.584 Yeah. We're building it all open source and we want to help other organizations use the tool in a similar, same way that we are. So the grant will help fund us to actually build out a lot of it and get it going, and actually make it available for others to use, and make it easier for others to use as it is right now.
RVB: 00:12:05.313 That makes a lot of sense. And what do you think about the future of the category, the future of graph databases in general? Any perspective on that?
JM: 00:12:16.739 Yeah. I mean, every year we go to GraphConnect and hear about the statistics on how many more industries have switched to graph databases. And I know I heard you speak once before about this becoming a standard database and not just a one-off or each-case database that people would use. This is becoming so much more mainstream that it just seems the most optimal way to move forward because you get the benefits of traditional databases as well as the enhanced powers of having a graph database, which is pretty incredible to have both of these cases in one.
JC: 00:12:59.768 And yeah, I see just the amount of cost savings that early-stage startup can get when they're in the see if the customers even want to use the tool and get it in early-user hands instead of building a full AI system to provide recommendations and train it on tons and tons of data. Really early-stage startups can get those features early on and not in a prototype way but a actual, functional scalable way.
RVB: 00:13:34.803 Absolutely. We were joking a little bit about artificial intelligence for humans, right [laughter]?
JM: 00:13:42.242 Exactly.
RVB: 00:13:43.916 Guys, thank you so much for taking the time to share that story with our listeners. We'll put a bunch of links to your work on the transcription on the podcast. And I'm sure lots of people will find you there. Thanks again and good luck with the rest of your project. It sounds like a fascinating ride.
JM: 00:14:09.105 Thanks again, Rik, for having us. We appreciate it.
JC: 00:14:12.229 Yeah. Thanks a lot.
RVB: 00:14:13.752 Yeah. Thanks, guys. Have a good one. Bye.
JM: 00:14:16.172 Bye.
JC: 00:14:16.785 Bye.
All the best