Appealing to young professionals has always been an ever changing task for social media giants and LinkedIn is no different.
With a global presence comes different challenges in each region and LinkedIn China faced its own issues when launching a service called ‘Chitu’.
The professional social media giant’s local arm in China saw an untapped market with young professionals and so the service aims to help jobseekers through a game-like user interface on desktop and mobile.
The service allows users to grow their networks via connections and direct messaging, as well as supporting opportunities for in-person meeting and networking events.
Numerous challenges were faced when deciding to create this service so they turned to Neo4J, a graph database that allows customers to connect data and built intelligent applications.
CBR’s James Nunns spoke to LinkedIn China’s Bin Dong, manager of development, Chitu.
JN: Why did you decide to use Neo4j? What other tools did you look at?
BD: “In terms of database choice, the first decision was whether or not we’d use a relational database. We decided not to, firstly because the relational schema would cause problems for fast iteration and we were also worried about possible migration and schema inconstancies.
“Secondly, complicated queries, like looking for second degree friends, are really hard to do in a relational database. I have more than 10 years of experience of using relational databases, so I know there is a better way to build a social network; I found graph databases are that more appropriate way. And after comparing several graph databases, I think Neo4j is the best choice for us here, according to its popularity, documentation and query language, Cypher.”
JN: What Neo4j tools are you using?
BD: “Just the Neo4j enterprise edition server. The programing interface is Cypher, and we’ve also implemented some Neo4j plug-ins for richer features.”
JN: What were the biggest challenges facing you? How did you overcome them?
BD: “Before Chitu’s launch, the biggest challenge was how to deliver a complicated social app before deadline! Neo4j helped a lot in terms of making our development process very efficient, and we also utilised some technologies to increase our speed.
“Queries can now be performed in record time, and we were able to launch the first version of our application in only four months.”
JN: How did you manage to successfully target young professionals?
BD: “The system is Chinese-only from a user interface and content point of view, it’s proving easy to consume by our market here, mainly young professionals with a strong educational background.
“Unlike LinkedIn, Chitu is mobile-only with a messaging system supporting photos, emojis, geo location and other advanced content. Chitu also supports knowledge sharing and monetisation through live presentations that the user can subscribe to on career development issues.”
JN: Dealing with the challenge of scale, how do you overcome this?
BD: “With a Neo4j cluster, we have one master for writing and other slaves for reading. When traffic increases dramatically, we scale the database simply by adding more slaves, because writing traffic isn’t ever going to be that large an issues for a social app.”
JN: What other technologies are you using and how do they impact how you work?
BD: “We use GO, Mongodb, Redis and Elastic Search, all new and fantastic technologies. Compared to other programming languages, GO is the best combination of robust, development efficiency, while Mongodb is very stable and fast, and used to store posts and chat messages.”
