Connecting the dots

I was very lucky in life?—?I got hold of computers back in the early 90's when they were this new thing. Especially back in Russia (actually, I think I got my first computer when it still was USSR). I never had too much interest in playing games (my parents would probably disagree), but had more interest in building stuff.

My first post-college (technically I started it in my sophomore year) job was at a large telecom company, where I was the only developer in the Value Added Services department: back in the early 2000's when monophonic ringtones and b/w icons were the cutting edge I was developing platforms to distribute those. Since it was a large telecom company there were also a lot of other platforms I had to integrate with?—?billing, provisioning, all sorts of gateways, positioning systems and many more. That’s when I realized that the actual power of modern computing is not as much in developing new stuff, but rather in creatively connecting already existing pieces.

Virtually every developer these days uses libraries developed by other developers, we use API’s that provide functionality we don’t need to spend time developing ourselves. When I was writing my thesis at college I had to write a language parser myself, right now there are several dozen Natural Language Processing frameworks available as open source.

Development is a lot about knowing what tools and libraries are available and creatively combining those. We started SCOOP just to combine CrunchBase and AngelList, because those were the tools I was using daily. It’s incredible how much data you can collect starting with the domain name of a company you’re looking at.

So far it’s been all about collecting information and displaying it, but there are also numerous tools and services available for processing this information, for analyzing it and generating insights. And that is the next step for SCOOP?—?we’ve started by connecting the 4 sources of data (CrunchBase, AngelList, FullContact, Compete). We’re going to dig even more information from those sources than we are currently displaying, and will take the next step and use some of the tools out there to analyze it.

There are a lot of ways to go from here?—?we just need to find the right dots to connect to our existing thing. But there are so many crazy sexy cool AI/machine learning and Natural Language Processing libraries and API’s out there that it’s sort of hard to resist the temptation and try not to extract some more insights from the data we can get our hands on.