I am on a quest to better understand how to collect, manage, analyze and report data. You can thank the line of work I’m in and that I am taking a Data Science class on Udemy.

In my last data series, I did a cursory analysis of Blockchain and artificial intelligence media coverage across traditional media. You can get this type of data from platforms like Trendkite, MuckRack or TechNews. It’s pretty simple to do–write a Boolean query into a search field, hit enter and there you go. You may have to go back to the Boolean and adjust it in order to isolate the exact coverage you are looking to explore. I call this type of analysis a market or media analysis since the data you are collecting is getting pulled directly from news articles.

It’s very similar to social listening where you would follow very similar steps–select the content sources, write your Boolean, maybe add a language or geography filter and then sit back and wait for the results. Sophisticated platforms like Crimson Hexagon do this very well, and even use machine learning to find and isolate the most relevant conversations in social channels, forums, and blogs (news articles too). They also cluster the data into actionable insights which can be used for just about anything–community engagement, crisis monitoring, customer care or building social campaigns.

I approached this analysis differently.

Instead of starting with a Boolean search to find/export conversations and then try and isolate the audience, I reversed it. I built an audience first, added them to audience panel in Crimson and then analyzed their conversations. As someone who has a lot of experience in working with B2B brands, I know the software developer audience is one that many of these companies want to understand so I started there.

The current panel consists of ~6K self-identified developers, programmers, and engineers. This was done mainly by combining bio and shared content data. For example, I used “developer OR engineer OR programmer” as bio terms and cross-referenced those accounts with shared topical data like hadoop, data science, and python. This method ensured that I was finding software developers, instead of commercial real-estate developers and/or chemical or civil engineers.

I ran the list through Audiense and further segmented the developers to understand what I was working with. I discovered the following sub-segments within the audience panel–IOS Developers, Microsoft MVPs that are using Azure, VMware vExperts, PHP Developers, Game Developers, Data Scientists and Security/Hackers.

This method allowed me to isolate the conversational data and drill-down on the topic areas that were top of mind for each sub-segment. The data showed that the top trending topics among this developer audience revolved around nine core areas–artificial intelligence (AI), big data, security, IoT, machine learning, blockchain, data science, devops and deep learning. I used these insights to then create filters within Crimson Hexagon and isolated the data even further within more of a historical context. I pulled data from 1/1/17 to 6/1/18–roughly 18 months which resulted in 16.5 million posts (or conversations) among the 6K software developers.

Here’s what I found:

Clearly, developers are all over artificial intelligence. In fact, AI leads the way in terms of total volume, followed by Big Data, Security, and IoT. You’ll notice that Machine Learning and Deep Learning have separate data points. They were purposely excluded in the AI analysis in order to get the most accurate view. Similarly, when analyzing Machine Learning, both AI and Deep Learning were excluded and so on.

TOPICAL ANALYSIS

The real value comes when analyzing topical data within the audience panel. For the purposes of this post, I only extracted the top media publishers that developers are reading and sharing based on five of the nine topics listed above–artificial intelligence, big data, security, blockchain and data science. The publishers are sorted based on total interactions from the developer panel. Higher interactions typically mean that the content resonated with those who are reading it.

The following data represents the top five media publishers that developers read and share when talking about and sharing news related to artificial intelligence.

The following data represents the top five media publishers that developers read/share when talking about and sharing news related to big data.

The following data represents the top five media publishers that developers read/share when talking about and sharing news related to security.

The following data represents the top five media publishers that developers read/share when talking about and sharing news related to blockchain.

The following data represents the top five media publishers that developers read/share when talking about and sharing news related to data science.

Oh, and a few bonus insights here. Developers prefer Spotify as a music service. In fact, they often share playlists of melodic heavy metal and EDM (had no idea there was such a thing) with their friends and colleagues, so they can also concentrate when coding. They also share a lot of long-form content from LinkedIn blogs and Medium too, much of it is content they wrote themselves.

SO WHAT DOES THIS DATA ACTUALLY MEAN AND HOW IS IT ACTIONABLE?

Well, it really depends on who you are and what you do. If you work in public relations and you want to reach software developers,  you can start to dissect the above media publications, build a media list and prioritize your media outreach. You can also start to align your messaging and narrative to one or more of the above topics, assuming it’s related. You may also consider submitting an executive byline to one of the publishers. Or, you can pay $1,200 to Forbes and join their technology council and post content that way.

If you work in digital marketing, you may consider a media buy or sponsorship on a few of these sites since you know that the developer audience is spending time there.

For those who work in social media or content marketing, you can build a custom audience and target content to this group with minimal paid investment.

THE REAL VALUE IN THIS DATA

The social data above accounts for 18 months of conversations, sharing and engagement. It’s great for historical context and to understand the trending conversations that are peaking and the ones that are on the decline. In order to make this data more useful though, I would consider listening to what the audience is saying in real-time and respond in real-time. This will help brands be more relevant among the people that matter. In fact, I see this approach as a strategic replacement of what many brands are still doing today–creating content calendars for weeks in advance, going through days of approval processes and scheduling posts using automated technology. There is little to no value in doing it this way anymore, especially when all the social content is linking purely to owned media (blogs, white papers, etc.) that offer very little value to the developer community at large.

I hope this quick analysis was valuable. I plan on doing these “Data Series” posts once per week in my quest to brainiac status. Help me get there.