Josh Livni has been making maps ever since he started getting lost in the wilderness. He works on the Google Crisis Response team, helping to make actionable information more accessible during times of disaster. Before joining Google, he ran a consulting company, integrating cartographic and statistical tools on the web. He cuts his own hair, likes his beers bitter, and his salsa spicy.
Q: How did you get into mapping/GIS?
A: Well, my wife thinks it’s because I can’t find my way around anywhere. Which is generally true: I don’t have a great sense of direction. Starting in high school I spent a lot of time in the wilderness, and maps were like magic to me as I figured out different routes. By the time I got to college I started to think that the integration of technology and maps was imminent and I really wanted to be part of it all.
After working at a streaming media startup during the first “.com” boom, I decided to make a career for myself that allowed me to focus on technology and the environment. I had a degree in environmental studies, and GIS seemed like a great way to mix some computer skills with figuring out how the earth works, and hopefully having a positive impact.
I started out volunteering at an environmental non-profit, where I taught myself how to use Esri and related proprietary software. From there, I slowly switched to using open source as I needed to programmatically handle bigger datasets, longer running processes, and webmaps. I never turned back and I’ve been working in geo now for over 12 years. I feel lucky to have found a profession that is so perfect for me and that I enjoy doing everyday.
Q: You recently transferred within Google from being a geo developer advocate to the Crisis Response team. What will your new duties include? Will you be doing more or less geo stuff than before?
A: There’s almost always some spatial component in getting useful information to people who have been affected during a disaster, so I’ll still be working with lots of geo data. But as always, there are many more ways to effectively communicate spatial data than simply placing it within the context of a map. My colleagues on the Crisis Response team have put a lot of thought into this (and many other areas), and I’m going to be focused on helping automate and scale more of our response processes to bring actionable information to those affected, more quickly, across the world. The exciting part here is mixing spatial content with other data, where the concept of geo goes beyond maps and cartography.
Q: I was an early user of your shpEscape tool, which loads shapefiles into fusion tables and now also converts shapefiles to GeoJSON and TopoJSON. I love shpEscape — it fills an important void. Will you continue to enhance and add functionality to shpEscape?
A: shpEscape was actually a weekend project of mine many years ago. The Fusion Tables API doesn’t accept shapefiles as an input, but it was gaining a following amongst non-GIS folks who weren’t sure what to do with this “shapefile” thing they had downloaded. The code was originally designed to be throwaway, and the site was never really advertised; I’m constantly surprised it’s still running, let alone used. But I’m glad you like it! When people tweet me if the queue seems stuck (usually when someone uploads a few very large complex files), I often say: ‘Yup, I should really think about working on that sometime.’ Maybe it’s that time now and I’ll check it out again: What enhancement would you like to see?
Q: How about the ability to handle all features of the “enhanced” KML output of Google My Tracks, which currently comes across as just two points (beginning and end) in most mapping applications?
Well shpEscape only accepts shapefiles as input. But adding more formats, including KML, is definitely a good idea. I’ve had in the back of my mind a total rewrite that turns it into a general interchange site for any spatial data for a while now.
You may also be interested to know the OGC KML Standards Working Group is currently discussing whether to put gx:Track into the upcoming 2.3 spec. While some of the gx: extensions to KML don’t make sense for 2D-only applications, or those without robust temporal visualizations, this one certainly does. If and when it becomes part of the KML namespace, I’m optimistic we’ll see more applications accepting it.
Q: Do you think the current open/closed source balance — within and without the geo industry — will change significantly in the near future? Will open source continue to gain market share?
A: There are a lot of different markets, and in some niche verticals open source may never gain traction. But overall yeah, I expect open source software will continue to play a bigger and bigger role in both geo and the rest of the software ecosystem.
Q: Heartbleed and the recently-discovered OpenSSL vulnerability have bolstered skepticism regarding one of the main advantages of open source. Does it make a difference that the code is available for review if nobody reviews it?
A: I’m not convinced that security is a priority for most developers. That is the real problem, and I think it affects both open source and proprietary software similarly. There are exceptions on both sides, but developers are mostly interested in features, like ease of use, or interop, or whatever. Getting stuff to work is hard, so most developers focus on just getting stuff working, which is why everything is broken.
Q: We define hipsters as people who think outside the box and often shun the mainstream (see visitor poll with 1106 responses). Would you consider yourself a hipster? How do you feel about the term hipster?
A: I think that poll has me nailed: I almost always prefer GeoJSON (my only complaint is no explicit winding order), and I would never refer to myself as a GeoHipster.
Q: Geohipster (and geohipsterism as a concept) is sometimes criticized for being exclusive and/or attempting to foster divisions within the industry. Or sometimes for being different for the sake of being different. You have published the source code for shpEscape. Did you do it to be different?
A: Cliques can be important for building depth within a specific community, but I lean towards breadth as being more useful (at least for me). When webmaps became popular, the neo-geography crowd purposely avoided a lot of knowledge from the GIS community. By passing up a lot of unnecessary complexity, one result was a tremendous upswell of simple and elegant tools, but also a lot of mistakes and miscommunication that still hinder us today.
Looking at the poll results it seems like people who responded to being geohipsters are bridging that divide more than coming up with anything particularly exclusive, which wouldn’t have been my initial definition of a hipster, but I think it’s good in this case.
As for publishing the source code for ShpEscape, I did it because open sourcing stuff like that just makes sense. I recall being a bit embarrassed by the code quality, but that wasn’t the point of the demo, nor is it a good excuse to keep it hidden. I doubt many people have actually tried to deploy it though; I’d like to go back to it one day with a fresh rewrite as a generic swiss army knife for transforming data, with a more reasonable architecture.
Q: You fly paragliders, which is something I always wanted to try. How did you get into that? Did you do that to be different? 🙂
A: No, I tried paragliding to see if it would be awesome. For as long as I remember I’ve had dreams of flying, and when I first heard about paragliding I thought it might be boring (sitting in some kind of chair didn’t compare to my superman-style vision). But a friend of mine convinced me to give it a shot, and I was totally hooked. In some ways, it’s more amazing than my childhood dreams were. Unfortunately I have not been up for a long time. As a personal goal, I’m going to go flying before I work on shpEscape. Sorry!