NODES 2019: Nine months on


Neo4j’s Lju Lazarevic describes how she and her colleagues created an immersive and rewarding online event when the global pandemic affected their plans for an in-person event.



Ljubica: Hello, everybody. Thank you for joining me today. I’m Ljubica Lazarevic, I’m a developer relations advocate at Neo4j, and as mentioned, we’re going to talk about our online conference NODES 2019.

So very quickly, who are we? We’re a graph database company. And I can hear some of you going, “What?” So what’s a graph database company? So we talk about graph databases, we’re talking graph as in graph theory in the branch of mathematics, and this talk’s about understanding how discrete entities are connected to each other.

So for example, if I’ve got two discrete entities, so Jane and car, we’d also want to understand in the graph what’s the relationship, what connects Jane and car. So we’ve got these two nodes that we’ll refer to as vertices in graph theory, and they are joined together by a relationship or an edge as it’s known in graph theory.

So we can now see that Jane owns a car, so we understand the connection between these entities. And the big difference with a graph database, say to a relational database, more traditional database is that we treat the relationship between our discrete entities of data as important as the data themselves, and we store all that together.

And this is really powerful if you want to understand networks, how people are connected to each other, how things propagate, influence, and that kind of thing. So that’s graph database in a nutshell. A little bit more about the company. So we’ve got a really active, vibrant community. We have over 150,000 subscribers to our weekly newsletter. This week in Neo4j, we have got over 5,000 active users on our forum, over 11,000 users on our user Slack channel.

These wonderful members of the community help each other out. So they’ll pick up questions that come up, they’re answering questions on things like stack overflow as well. So they’re really active in the community and helping everybody along in their growth journey. They also produce a lot of blog posts, and this can be anything about their experiences of working with a graph database, to how would they work with their favorite framework, how would they integrate that with the database, to creating code snippets, drivers for different languages to connect to the database, and a lot of other content as well.

And they’re also very active. So they’ll come and join in our meetups, maybe they’ll go into other user group meetups, so let’s say Java user group or a Data Science group, and they’ll talk about their experience with database. And they attend conferences, both external conferences, but also our annual conference which we have GraphConnect.

And mainly, a lot of our conferences that we have and events that we have are in person. This year is slightly different. We have more events, obviously online given the circumstances, but last year, certainly, it was mainly in person. So why did we decide to have an online event? Unfortunately, the decision was made to postpone GraphConnect, our annual conference for 2019.

And we had this situation where we had a lot of members of the community who had done a lot of amazing projects that they’ve worked on, the experiences that they had, and we wanted to give them an opportunity to showcase what they had done. And the challenge that you have with a conference is in-person conference, it’s not always possible to attend that conference.

It takes a lot of time especially you have to travel there, so it’s not just the time of the conference, it’s also getting there, getting back home again, the costs with hotel, with transports, with other incidentals that come up. And if you think about what comes out of the conference, either virtual or online, you’re still going to capture that content.

So all of these sessions are going to be recorded, and they’re going to be available for people to watch in the future, so you’re still fulfilling that need. And by having an online conference, we’re able to allow our community to talk about their experiences. We allow them to do it in a way that was low cost and more convenient, and people still get to enjoy that fantastic content afterwards.

So the Neo4j Online Developer Expo and Summit was born, and you can probably see where we took the inspiration for the name. So let’s talk a little bit about the run-up to the event. So first of all, a blog post was published announcing that we were going to be doing this online event.

And in there we talked about some of the things that we would be doing, the run-up towards the event, but also we announced our call for papers. So this run over a two-month period, and we had so many high-quality, amazing submissions that we’ve took the decision from what was previously planned to be a three-track event, and we turned it into a five-track, one-day event.

And this straightaway demonstrates the advantage of having an online event. It’s really easy to make these kinds of changes. You sort of think about if this was an in-person conference, how you may not necessarily be able to hire another room for the track. Maybe your budget doesn’t allow for it. But straightaway, as an online conference, you have the opportunity to scale, and we were able to say yes to more excellent talks.

Another event as well being organized by my colleague Corinne, reviewing parties. So maybe less applicable at this moment in time, but the idea of these were to encourage our community to host both public and private events. So private events, the nearest and dearest colleagues and friends, public events that were open to everybody.

And these were hosted worldwide. This is a great opportunity to get together with like-minded people, and have that experience of being able to talk to people of the talks you’ve seen, that you’ve attended, any questions that you have even though it’s a virtual session. And last but not least, we had the Global GraphHack. So normally, with our annual event, GraphConnect, about two days before the event we host a meetup…so, host a hackathon.

And this would either be people local to the area or people who’d already come over for the conference, and they’d compete in this hackathon, and then at the end of the day, we would announce the winners. So we were keen to do a hackathon, but we made it slightly different. And what we did here is we encouraged global participation. So we encouraged people to join together either with local teams or have a team consisted of global participants, or even people who wanted to take part, we’d help put them together into teams.

And it was very interactive. You had members of community downloading the various submissions, playing with them, providing comments, providing reviews, and it was really, really hard to judge the winners in the end. And the thing to emphasize here, the theme of the hackathon was sort of improving the graph community. And it wasn’t just about expertise, it wasn’t a case of, you know, the person who had the best graph skills was going to win.

It was all about novel and thoughtful submissions. And one of the winning teams came in with the experience of learning about graph. So it was a really great, great session that really helped the community. Some of the submissions that came through last year are still being used and improved and iterated. There were a few bumps on the road as we’re going through this journey.

So one of them was switching over from Crowdcast to BigMarker. So originally, we were planning to do the session with Crowdcast. Unfortunately, it wasn’t so great at handling multiple tracks. And something that we really liked about BigMarker was, as well as the multiple tracks, it allowed us to work more per session.

So if one session for whatever reason was overrunning, we could start the next session on time, and it wasn’t a problem. If something needed to run out because there was an extended Q&A, that was not a problem. Unfortunately, we made this decision about two weeks before the event, and there was some confusion.

So on the day, a lot of people were turning up to Crowdcast and weren’t sure why they couldn’t see the session. So we had to spend some effort triaging that and being able to get people over into BigMarker. And another challenge as well, not unique to us, we had speakers pull out including people pulling out a day or two before the event. Fortunately, we had other speakers who were able to step up and fill those slots.

So on to the day itself. We were in Malmo in Sweden, so that is where our engineering department is based, and that was the previous home of HQ. And we rented a studio in the area, and at 8:00 local time, we were there in the studio, and we were doing the setup, delegation of tasks, who was going to be monitoring which tracks and so forth.

And then at 2 p.m. that would have been 8 a.m. Eastern time, we started off the event with the keynote from our CEO, Emil Eifrem, and then at 4 p.m. local time, so 10:00 Eastern time, we started all five tracks, and this went on until 11:00 in the evening, local time. It was a long day, but it was great fun, and we were absolutely blown away at the amazing talks that our community provided.

So now let’s talk about some more of the elements. And for us, it was really important to keep it interactive. And it’s really easy to have the interactiveness when it’s an in-person conference. You bump into people you’ve met before, maybe a talk’s been real interesting, and you want to go and talk to presenter more about it. You have the hallway track.

So we were quite keen to think about what could we do to try and keep it interactive. So some of the things that we did. We had an ask me anything session with our CEO, Emil. And what happened? We encouraged the participants, the audience that were there, to submit their questions in the chat, in the platform, whilst the main keynote was taking place.

And then we collected up all of the questions that came in, we had them up on the tablet, and then Emil was going through and answering them. And I think that’s really powerful. That’s a really great way for the community to be able to engage with your company, especially if it’s the CEO who’s going through and answering those questions, and I thought that was very powerful. Another thing we did as well.

So we talked about how the viewing parties were being organized, we really loved the photos that we were getting of all the viewing parties that were taking place around the world and people sending us their pictures. And this is again a great way to be able to interact and have a bit of that conference feeling and be able to talk about the talks. And last but not least, we have the Hunger Games.

So this is our Director of Developer Relations, Michael Hunger. I’ve been told that the naming of this challenge had nothing to do with the famous book and movie series. But effectively, what the Hunger Games was it consisted of three questions. So we asked all of our speakers if they could think about three questions that were based on that presentation, so an easy question, a medium question, and a hard question.

And then these questions would be set at the end of each session, and there were prizes to be had. And this is a really interesting mechanism because not only do you encourage people to focus on the session because there’s going to be some questions to answer, but it can also encourage more participation, so maybe there are more questions now that you want to ask.

You have the degree of competitiveness as well. And having questions like this at event is a really great example of something that works really well for a virtual event that might be harder to do in person, so this is something great to explore. This was fun and successful, and I loved the Hunger Games.

So it was a great experience, and absolutely, when we did this, we took a step back, and we thought about what happened, what worked well, what perhaps we could do better. And to summarize some of the points here, so the first one, make it clear it’s a free event.

Certainly now where there’s a lot of events going on, some are free, some aren’t. And we feel that maybe people were put off slightly, or people were reusing the same credentials for logging because it wasn’t clear that it was a free event. So promote it, make it clear if it’s a free event that it’s a free event. Another thing that we thought would work really well is creating a how-to conference guide for your participants.

So maybe it’s not always clear where you ask questions or where you would chat. Perhaps they need a bit of help of how they’re going to interact or do the Hunger Games, for example. So something we’re quite keen to explore is to do a five-minute how-to guide to get people comfortable and get the maximum out of the conference experience. Have more ways to connect with the community or for the communities connect to each other during the conference.

So we had the chats in the actual sessions while some we’re going on, but once the session ended, that was it, the chat window disappeared, and there was no element of continuity. So something we want to look in to improve, and certainly what I’ve seen with a lot of online conferences like this one, for example, is that there is a Slack channel to enable that.

Check to see if your speaker has any need for mitigation options. Does your speaker have a good internet connection? Do they have a quiet room where they can present, especially now where you have remote working and a lot of homeschooling going on, is there a more preferred time for the speaker where they know it’s going to be quiet and adjust the schedule accordingly?

Maybe they’ve only got one screen, which is a laptop screen. So perhaps it’s a good idea to get their phone number, so you can relay any questions coming in, and they don’t have to flick between their screen and the question screen. And another one, allow a bigger gap between sessions. So we had 45-minute sessions, and they came one after the other. And the idea was a 30-minute talk, 10 minute for Q&A, and then 5 minutes for Hunger Games.

But sometimes the talks overran which is great because BigMarker could handle it, but you maybe want a bit of break between the talks for comfort, get some fresh air, or just to give you enough time to prepare yourself for the next session. And last but not least, don’t try to do it all. We’re a relatively small team, and we had one person per track, sort of monitoring it and so forth.

And it became quite clear that it could be a bit hard especially if you were moderating track, but you were also going to be doing a talking. So you didn’t necessarily have an opportunity to do any preparation, or just get a bit of a breather. So again, either get more people involved to help you, or maybe reduce the scope slightly if that’s not possible. So it was a very successful event.

We enjoyed it a lot, and we had over 4,600 registrants with over 1,600 live attendees on the day itself. We had 52 sessions plus the keynote, and we had presenters from 15 countries. It was massively successful. It was well received. We had a lot of amazing feedback, and we kept being asked, “So when’s the next NODE? When is the next NODES?”

Well, I’m delighted to tell you that NODES 2020 is coming. It’ll be in October on the 20th. It’s a free event. There we go, I’m applying the first lesson there. And if you’d like to know more about NODES, if you would like to know about getting updates, if you would like to register, if you would like to learn about graph databases, apply it in your use case and then submit a talk about it, you can do all of that at

And finally, if you have any questions about this or something I’ve not covered, or you want to know about graphs, then feel free to contact me. So you’ve got my email address there, or you can send me a tweet or a DM on Twitter. So thank you very much for joining me in this session.

Tamao: Thank you so much, that was amazing.

Thank you.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.