Monthly Archives: October 2014

Amy Smith: “One of the things I love about Maptime is that it’s open to all skill levels and backgrounds”

Amy Smith
Amy Smith

Amy Smith is a Geospatial Data and Technology Specialist with Fehr & Peers in San Francisco. She’s had some great opportunities working with geographic information systems in a variety of fields, including environmental studies, satellite imagery analysis, water resources, and transportation planning. Amy currently spends her days working with an amazing group of people focused on improving transportation in our communities. In her free time she enjoys exploring the hills of San Francisco.

Amy was interviewed for GeoHipster by Christina Boggs.

Q: A few years back we used to work together but I don’t actually remember how you got into GIS. You have a master’s in it right?

 A: I do! I have a Master’s in Geography and Regional Studies. I got into GIS through a chance encounter with a geography professor whom I passed in a hallway on campus. Somehow we started talking about geography. I was undeclared at the time. I was trying to decide between GIS and intro to computer science for a general requirement. He told me a bit about GIS and geography, and that really won me over. Who knows, I could have been a computer scientist if I hadn’t met him!

Q: You didn’t leave computers entirely though, you’re pretty slick with code… In fact, you’re a great promoter of Python. Which came first – GIS or coding?

A: My first programming/scripting language was Matlab. I learned it while I was working on my master’s studying space-based synthetic aperture radar data in the Florida Everglades. Through learning Matlab, I learned the basics of programming logic. When I started using desktop GIS every day for work, it got me thinking about ways I could be using programming for spatial analysis, which led me down the path to Python. Since then, I use it almost every day, and not just for spatial analysis.

Q: What other tasks do you use Python for?

A: Lately I’ve been using it to prepare transit data for travel demand models. Since many of the inputs of the models are text-based, Python lends itself well to these types of tasks. It can also come in handy for automating things you’d rather not do manually. For example, I had an Excel spreadsheet with multiple worksheets that needed to be saved as individual CSVs. Instead of exporting them one by one, I wrote a script to iterate through each worksheet and save it as a CSV. Kind of a mundane example, but it’s this type of thing that I think can save lots of time at the end of the day.

Q: Speaking of time, you did a transportation study and saved time by scripting some node-based analysis of road segments to bicycle accident occurrences. I saw your talk at the ESRI UC where you talked about becoming one of the points. Do you know if that study has been reviewed by any of the traffic safety folks out in your area, has it helped any?

A: That was one of the first projects where I got a chance to develop a custom script tool for ArcGIS. The tool uses a roadway network and collision data to pinpoint high incident collision areas that might need attention. The tool was applied most recently by Placer County here in California to run a collision analysis of their entire county-maintained roadway network, which used to be a manual review process. They used some of the results to apply for grants and received several grants funding highway safety projects. Another benefit of the tool is that the county can continue to use it with new data in their safety programs.

Q: You’ve taught workshops on Python and even done some online workshops. Do you have any more in the future, or are you branching out to something different?

A: I’m planning some internal Python training here at Fehr & Peers for our planners and engineers who’d like to learn more about it. I’m always happy to talk with others about Python, so I hope there are more opportunities out there for workshops. I’m still learning too, so I’m always on the lookout for workshops and meetups others are hosting. In terms of branching out, I’ve recently been diving into JavaScript. There’s a library I’ve been learning about called D3 that has some great spatial as well as non-spatial capabilities. I’m still in the “stumbling through it” phase, but luckily there’s a great user community online and here in the Bay area that’s eager to share knowledge.

Q: A few months back you attended my first #maptimeSF with me; now that you’ve moved out to San Francisco I see you get to go to #maptimeSF more often. For someone who is thinking about attending their first Maptime, how do you think it helped you as an advanced GISer?

A: Maptime is a meetup that’s happening in many cities around the world where folks can get together, learn about maps, make maps, talk about maps, or maybe just hang out with friends. One of the things I love about Maptime is that it’s open to all skill levels and backgrounds. People are encouraged to ask questions and learn from each other. It’s a very welcoming environment. I’ve learned a lot about how others outside of my industry are using geospatial data and technologies. It’s also encouraging to see a thriving interest and enthusiasm for maps.

Q: Hearing about your work in transportation is really interesting. The water side still misses you. What are you up to at Fehr & Peers? Any interesting projects you can share with us?

A: I have so many great memories from my time with the Department of Water Resources in West Sacramento — it’s where I really started to get my feet wet (pun intended) with Python! It’s also where I learned to drive a boat. I definitely miss the field work collecting bathymetry data in the Sacramento-San Joaquin delta (picture below for proof), and of course the people too!

Amy Smith collecting bathymetry data in the Sacramento-San Joaquin delta
Amy Smith collecting bathymetry data in the Sacramento-San Joaquin delta

 One of the great things about GIS is that it’s applicable in many different industries. Transportation planning has a lot of great uses for GIS too. One of the more recent projects I worked on focused on improving cyclist and pedestrian access to transit stations. The project had a large data organization component that involved gathering available spatial data and organizing it in a consistent way so that we could use it in a series of network analyses. We looked at some of the ways that a well-connected network might help improve access to transit, making it easier for people to walk and bike to stations. I’m currently working on a project, also transit-related, that involves improving transit in an area that doesn’t have a lot of existing transit. It can be a challenge to anticipate how new facilities will affect travel in an area if you don’t have many observations on how people are currently using transit. In this case, we’re identifying places that have developed transit networks and that share similar characteristics with the study area that’s considering improving or expanding their transit system. Both of these projects are very much rooted in spatial analysis, but also require local knowledge. Another fun part of my job is getting to know new areas and talking with people to learn about qualities specific to their region that might not be obvious from just looking at the data.

Q: If people are looking to check out some of your cool stuff, where can you be found online?

A: I tweet about spatial topics, transportation, and my endless appetite for spinach @wolfmapper.

Q: Geohipster Amy Smith is awesome! How do you feel about being part of spreading the geohipster gospel?

A: I’m a big fan of GeoHipster! I was trying to disguise myself a bit by using a seriffed font, but I think you found me out anyway. 🙂

Q: Speaking of transportation, I’ve got to wrap this interview up so I can cycle to work. Is there anything else you would like to share with #geohipster readers?

A: I recently learned about a spatial data format called topoJSON that’s one of my new favorite things. I found out about it at a recent Maptime on D3 and have been reading more about it on Mike Bostock’s wiki. Also, I’ll be helping host a webinar on transit planning with Code for America next month. Tune in if you’re interested!

Happy cycling!

Ed Freyfogle: “Every startup is vulnerable, that’s what makes it exciting”

Ed Freyfogle
Ed Freyfogle

Ed Freyfogle is a German/American entrepreneur living in London. He is one of the founders of Lokku, makers of the OpenCage Geocoder.

Ed was interviewed for GeoHipster by Atanas Entchev.

Q: You are the only geo person I know with an MBA from MIT. By contrast, I know people who wish they had an MBA from an Ivy League school so they could get out of geo. So what is your story? How did you get into geo?

A: It’s a bit of a long story, so bear with me.

I guess like most people “in the industry” I’ve always liked maps, as a kid was drawing maps, all that kind of stuff. Before MIT I had worked  as a software developer at Yahoo Germany during the first internet wave of the late nineties. It was a great case of being in the right place at the right time. I learned an immense amount. After five years there, with all its amazing ups and then the downs of the crash in 2001, it was time for something new. Also, while I really enjoyed programming I also wanted to learn the business side. So I got an MBA at MIT and thought a lot about what I wanted to do, and where I wanted to do it (I mainly grew up in the US). My conclusion was that I had really enjoyed Yahoo when it was small and had the startup feel; when I joined the Munich office was 15 people and I got to work on pretty much every system. I also concluded I had really enjoyed living in Europe. So in 2005 I moved to London in the hopes of finding a startup to join. Back then the scene was microscopic compared to today, I couldn’t find a startup where I liked the people, the idea, my role, etc. So in the end I started my own company, Lokku, along with another ex-Yahoo, and we’re still thriving today.

Those that know their geo history will recall that 2005 was the year Google Maps came out, shortly followed by Housing Maps, the first “mash-up” to put pins on a map. Heady times! Lokku’s first product, and still our biggest, was a real estate search engine called Nestoria. Initially we were just for London, today we’re in nine countries. A friend of mine from Yahoo, Mikel Maron (who later went on to start Humanitarian OpenStreetMap Team), knew about maps, and was advising us, and that’s how I got into the London OpenStreetMap scene, going to a few of the pub meet-ups. I’m proud to say we sponsored the very first State of the Map conference back in 2006 (and have sponsored many more since). In 2006 there was a mapping party on the Isle of Wight, and afterwards we made tiles and started using them on Nestoria if someone searched for a home on the Isle of Wight. I’m pretty sure this was the first ever commercial use of OSM in a consumer service.

A few years later we launched Nestoria India and Nestoria Brazil. To do that we needed geodata for those countries. I have the sense most readers of this blog focus on the US or Canada, and I have no doubt there are impressive technical and licensing challenges around getting US geodata, but if you want to experience some grade-A bureaucracy I invite you, dear reader, to try to purchase Indian geodata. It was impossible, at least for a tiny start-up like ours. So basically we had to launch using only OpenStreetMap as our geodata source. I can not pretend OSM in India is perfect, it still lacks coverage. But then of course so does any other geo datasource given how rapidly things are changing in India. Anyway, OSM was good enough, and now Nestoria India is one of our fastest growing markets. This is classic “Innovator’s Dilemma” stuff, the new technology is “good enough” for experimentation, and then all of a sudden it’s good enough for day to day stuff, and then all of a sudden it’s the norm, and the existing models with their old cost structure can’t compete.

So we thought about that and thought about whether there is a way we could help make that happen, and the result was last year we launched our brand OpenCage Data. Our hypothesis was that companies want to use OSM, but right now it’s too chaotic for them – the documentation isn’t always great, the way to learn is to get on mailing lists, the toolset around OSM is evolving very rapidly — and that all of this puts off companies who are used to more stability. We talked with lots of people, anyone who might have any possible use case for geodata, but especially people outside of the existing market. Companies put off by the cost, or not even really aware of how using geodata would help them. We learned a lot: OSM is not on most people’s radar yet. The thing that came up again and again was geocoding. So this summer we launched the OpenCage geocoder. We try to differentiate on simplicity/ease of use, by aggregating many different open geodata sources, and by then annotating our results with things developers would find useful. It’s early days, but we’re learning a lot and getting good feedback.

Finally, over the years we’ve been approached by a lot of different people asking for advice, help. We never had a good framework to channel that so we usually said no. But over the last year we’ve started seed investing in some of these ideas. We provide a bit of money, but also advice, connections, etc. It has to be in a category we have expertise in and one of those is geo, so now we’re involved in a few different geo startups.

Q: You are involved with several startups, and you run the #geomob London event. How do you manage to juggle so many different things?

A: Well some of the things I’m a driver, others it’s just as an investor / advisor, but yeah, there is plenty to keep me busy — also I have two small children, so there’s not a lot of down time.

#geomob is a regular event we run. It’s a lot of fun. It is amazing how many cool things are happening in geo and location-based services in London (the city where OpenStreetMap was invented). We try to create a forum to show off that innovation. Every few months we have an evening event where five or six different speakers get 15 minutes each to talk about their project. And it’s not all startups — we usually have a good mix of startups, hobbyists, academics, and the occasional megacorp. Afterwards we all go to the pub and have a few beers. We see lots of wacky ideas and experiments. And some of those crazy ideas turn into great things. Some of the speakers are polished, some aren’t. It’s a healthy mix from all across the geo spectrum.

Our next event is November 4th, if you’re in London, join us.

Q: I am intrigued by the business side of a geo startup. I watched your APICon 2014 presentation where you talk about OpenCage Geocoder — your latest startup. Your MBA background comes through strongly in that presentation. You are also very open about your business strategy. Doesn’t that make the business vulnerable?

A: Every startup is vulnerable, that’s what makes it exciting. No risk, no fun.

Lots of people who work in big companies or organizations perhaps don’t appreciate that with a startup the main challenge is creating momentum from a standing start. You start with literally nothing. And then you have to make it happen. No one comes to you. You have to create the momentum.

In general it is clear there is massive societal benefit to open data. But it’s not yet clear if all of that benefit goes to the end consumer and is just a cost society needs to shoulder (i.e., through taxes so that government services release all their data), or if there is role for private companies. Our bet is that there is. Anyone who has feedback on what we’re up to we’d love to hear from.

BTW, I’m choosing to take your “You MBA background comes through strongly in that presentation” as a compliment.

Q: In the same talk you refer to “Berlin, Berlin” as an error of redundancy. But do you know that there is a Berlin in New Jersey, another in Maryland, and yet another in Connecticut? Maybe “Berlin, Berlin” does make sense after all? Or perhaps addressing cannot — or should not — be standardized globally?

A: For those who aren’t familiar with the problem here’s a brief description. I was in Berlin, Germany, I tweeted, and Twitter showed my location as “Berlin, Berlin” (i.e., Berlin the city, in Berlin the state). Of course there are multiple Berlins, but Twitter has the coordinates from my phone. There is no ambiguity. They know I am in Berlin, capital city of Germany, yet they choose to show the location in a way that makes no sense to a local.

Absolutely addressing can not be standardized globally. It’s too late, there are almost as many formats as there are countries. That’s part of the rich tapestry of the human experience. Software should be able to solve the problem and present the location in the way a local considers normal.  I blogged about this and how we’re solving it on the OpenCage geocoder. Our solution is open source and we’d love everyone’s help. Pull requests here please.

Q: On addressing: I like What3Words and what they do, but how realistic is it to expect the whole world (including non-English speaking regions) to embrace an entirely new spatial reference system? Do you think this will happen before or after the US adopts the metric system?

A: Many industries and contexts in the US have adopted the metric system, as you’ll know if you buy a 2-liter of Coke. But I take your point that it isn’t the norm in most consumers’ heads. But so what? That doesn’t stop the rest of the world from using it to get their tasks done.

And it’s the same with a solution like What3Words. It is not an immediately compelling solution in a place like London, which is well addressed and has highly accurate postcodes. But if you’ve ever been to a meeting in India you will concede that there are parts of the world where addressing can only be described as a disaster. There are no addresses. You are navigating by landmark and frankly it is hugely painful. Not just for me the tourist, but for the locals as well. Those parts of the world need a better solution, and it needs to be one that is simple enough for the average person. That solution is not long/lat. I think it might be What3Words.

As a product person, What3Words is great in its attempt to make something complicated simple. I recently watched this excellent talk by Vladimir Agafonkin, maker of the mapping library Leaflet, on simplicity, and how it is needed in geo. It’s too early to say if What3Words will succeed, but I love that they are innovating by being simple. As an investor What3Words falls squarely in that category of most people dismiss it as crazy, but it just might not be, and if it succeeds it will be on a massive scale.

But if some parts of the world want to keep doing things the hard way or say measuring temperature in Fahrenheit, that’s cool. It’s a big world and there’s usually more than one way to do it.

Q: SplashMaps is another one of your business ventures, and the most hipstery one, IMHO. How did you come up with the idea? How is the business doing? Are the maps selling?

A: We’re just investors in SplashMaps, full credit goes to David and the team. But I agree with you, it is hip. It’s an amazing product, a customisable fabric map, perfect for all sorts of outdoor uses where you’ll get wet, muddy, sweaty, etc. So yes, in 2014 I’ve invested in a company that makes physical maps, which I guess is a little contrarian. I wrote about all our reasons for investing on our blog.

Good news for all the hipsters out there who can’t wait to get their hands on one (did I mention Christmas is coming?); in the very near future SplashMaps will be available globally, to date they’ve only been in Great Britian. If you know a geohipster who needs a gift, you’re going to have a tough time beating a SplashMap.

For me, SplashMaps is exciting because it’s a great example of the kind of innovation that’s possible when the barriers around access to and cost of geodata go away. I’m also more intrigued lately by taking digital products and bringing them back to analog as a way to create value. You can go on a tough hike using your digital map. But people want tangible artifacts they can hang memories and stories on. Everyone in the geo industry can remember the pleasure of gathering around an atlas, looking at far away tropical islands, sliding your finger along a journey you took. This tangible experience is a basic human urge that digital doesn’t meet, and one that SplashMaps taps into.

Q: I understand you moved out of Shoreditch, which is London’s counterpart to Williamsburg in Brooklyn. What’s up with that?

A: Don’t worry, I now live in the Barbican in central London. Shoreditch’s not far, but it’s gone a bit too upmarket. Anyway, I think of myself not so much as a hipster as a digital brutalist, so the Barbican’s a better match for me, my wife, and our two kids. All that said, these days the cool kids are all in Moabit and Wedding, so I’m working on convincing my wife it’s her idea that we move to Berlin. Let’s see.

Q: We haven’t talked about humor on GeoHipster, which I realize is a serious lapse. Let’s fix that. You were a humor columnist for the MIT student newspaper. Do you have a joke I haven’t heard?

A: Wow, you’ve done your research. Yeah, I used to write an anonymous advice column for the school newspaper called “Ask Alfred” in which I pretended to be a greedy and lecherous version of Alfred P. Sloan, the business school’s namesake (CEO of GM, often credited with inventing the modern corporation). In hindsight it was an attempt to poke fun at the divergence between the high-minded ideals espoused by the school and the profit-driven reality (greed, if you will) of the industries most MBAs go into. On the other hand though, I do think in the open geo world, particularly here in Europe, there’s a tendency to err too far in the other direction — thinking everything should be free all the time, all code needs to open source, we all need to be motivated by altruism all the time. As anyone trying to pay rent in central London will tell you, goodwill alone will not get you far.

I’m not sure what the joke is here. These days my comedy is more situational and slapstick.

Q: Thank you so much for the interview. Any parting words for the GeoHipster readers?

A: In the geo industry (if I can say so as an outsider) we all get that something major has changed with the rise of the smartphone; all of a sudden we all carry around a supercomputer that knows exactly where we are at all times. But I don’t think anyone grasps how radically this will change everything. For everyone, but particularly for the geo industry. We are at the start of an amazing ride. Anyone who’s up for the trip I would love to meet with. So if you’re London come say hello, we’ll go grab a pint. If that makes me a geohipster, so be it.

Brian Timoney: “The ‘G’ stands very much for ‘geospatial’ and not ‘GIS'”

Brian Timoney
Brian Timoney

Brian (@briantimoney) is an information consultant based in Denver, Colorado. With 15 years experience primarily in the geospatial sector, he has worked in a variety of sectors including energy, defense, and local government. Brian speaks both English and Spanish with a Philadelphia accent, and is a US Marine Corps veteran.

Brian was interviewed for GeoHipster by Mike Dolbow and Atanas Entchev.

AE: You blog and tweet about geo, data visualization, and business analytics. Which of those is of the most interest to you nowadays and why? Is spatial special?

BT: While analytics are hot (if the prevalence of the term “data science” is any indicator), the usefulness of raw insight often hinges on a visualization that is both accessible and meaningful. As for where Geo fits in, the popularity of map-based listicles e.g. “38 Maps that Explain the Global Economy” either indicates a serious, ongoing demand, or the delight of the otherwise well-educated to engage in geographic thinking for the first time since middle school.

MD: Your blog post summing up findings on How the Public Actually Uses Local Government Web Maps is so succinct and forehead-slapping that I find myself constantly referring customers to it. Do you see any evidence that the tide is turning in local government web maps, moving from overcomplicated user interfaces to simpler designs?

BT: Anecdotally I know of some projects in Europe that were directly influenced by those articles, and some vendors have engaged with the ideas in those articles. Yet we still see local government opting for the visual grammar of desktop GIS because it both feels familiar and risk-free. Or put another way, government websites aren’t punished for the users who leave because of a crappy user experience.

But what is significantly turning the tide is mobile. My new favorite quote is from CartoDB’s Javier de la Torre who said the future of geo “isn’t an application with 100 buttons, but hundreds of apps with one button.”

MD: Your series on Why Map Portals Don’t Work expands on your observations, taking a deeper dive on the subject of simplicity in web mapping. If you had to pick one part of that series for customers to focus on, which one would it be and why?

BT: My kingdom for an auto-complete search text box! Google has made text-based search such an intuitive part of using the web that maps that leave the user no choice but to interact with map elements — pan, zoom, etc. — make discovering user intent much more difficult than it should be. And make no mistake, users end up alienated.

MD: You must feel some validation when you see Vladimir Agafonkin’s recent FOSS4G talk on how simplicity will save GIS. What do you think this means for the future of web mapping?

BT: If you see the world as a Pareto 80-20 proposition, then you could make the case that the history of web mapping interfaces is one of bloating the map out to solve the last 20% of use cases. Vlad’s exclusive focus on the core 80% of map functionality is what has made Leaflet so successful, especially outside the traditional GIS boundaries. He was the star attraction at the recent JS.geo, and his story highlights the need to keep our industry open to outsiders. Just to be clear, there are plenty of workflows out there where an OpenLayers or Google Maps or Esri Javascript API makes the most sense. But Vlad’s commitment to both Leaflet as an open source project and iterating only on a core subset of functionality  has served everyone very well.

AE: FOSS4G 2014 generated a lot of buzz and excitement, but is open source making serious inroads in the geo market space? Esri still rules the geo desktop. Microsoft is still king of the PC. Do you see this changing any time soon? Or will the desktop decrease in market share to a point of irrelevance?

BT: For me, the big takeaway from this year’s FOSS4G was that the “G” stands very much for “geospatial” and not “GIS”. I think back to the 2007 FOSS4G show in Victoria where there was much more seeing the market as the GIS market dominated by ESRI. Today, the opportunities run so much broader and deeper. Take an outfit like the Climate Corporation, who gave a FOSS4G presentation on doing geo things using ElasticSearch: they were an open data/big data/analytics startup that were bought for $1.1 billion. They had a specific operational need — spatial search — and I seriously doubt they spent 15 minutes thinking of it as a “GIS” problem but rather a very specific type of indexing challenge.

When you say “Esri still rules the geo desktop” I think that would be better expressed as “…the GIS desktop”. But what about people doing spatial things elsewhere? People are doing geo things in R Studio. People are doing geo things in IPython Notebooks. To see GIS as having a foregone monopoly on spatial analysis and mapmaking is to miss the much larger picture. Unfortunately, some of the biggest losers in this changing landscape are current students unknowingly suffering from a lazy Geography curriculum that offers little in the way of spatial reasoning and data fluency but instead only a mediocre grasp of a particularly byzantine desktop interface.

AE: I am reminded of this ancient Persian proverb: “The dogs bark, but the caravan goes on.” Microsoft bought Minecraft to keep the caravan going. Should Esri buy SimCity? Because, you know, geodesign…

BT: Everything I know about Geodesign I’ve learned from James Fee’s Spatially Adjusted blog. I have nothing further to add.

AE: Do you consider yourself a (geo)hipster? You are a skier. Is skiing a hipstery pastime? (Or is it only if your heels are disconnected, as Mike Dolbow suggests?)

BT: The hippest thing I’ve ever done was switch from pleated khakis to flat-front khakis.

I also shave every morning and tuck in my shirt during business hours.

But if hipsterism is essentially about alternative status hierarchies, then count me in. Ever since Boundless named me the 26th most influential Geospatial tweeter, I’ve been looking for a new lunch table to sit at.

While I ski, it’s only in the context of enjoying the finest in chairlift technology, for which I happily pay a pretty penny. Mike is correct: in Colorado, “real skiers” are the tele-markers who ascend mountains using skins and free-heel down in knee-deep powder off-piste. To them I say “we will each go to our respective graves with very different ideas of ‘fun’”.

MD: What’s up with the “Geospatial Amateurs”? Has your intentional irony achieved the desired result? Can those meetups only take place where weed is legal?

BT: Geospatial Amateurs was the brainchild of Peter Batty (@pmbatty) and Nate Irwin (@nateirwin) and was very much informed by experiences with previous iterations of developer groups and meetups where well-intentioned sponsorship by vendors ended up creating environments that weren’t really what was originally envisioned. By putting “Amateurs” in the title, we accomplish two important goals: it’s a signal to vendors that this isn’t really anything you need to bother with, while communicating to curious outsiders that while it might be a bit nerdy, it’s not self-serious and insular nerdery.

As for weed, clearly you fall into the pattern of most of my East Coast friends for whom it holds an exotic allure akin to a 19-year-old frat boy pondering topless beaches in Europe. I’m within a 5-minute walk of three retail outlets and yet I guarantee you have given it more thought than me over the past few months. But to reassure you, the Geo Amateurs seem much more into craft beer.

AE: Thank you for the interview. Any parting words for the GeoHipster readers? Is spatial special?

BT:  Stop reading blogs during working hours.

Oh, you’re waiting for your buffer intersection to finish?

Cool — you’re still billable. Carry on.