SimpleGeo Launches | Hacker News Reader

SimpleGeo Launches

simplegeo.com

44 pointskingkilr29 comments

Hi, I'm the CTO and cofounder of SimpleGeo. Many of your points about sparse documentation, website, etc. are fair. We're about 3 hours old so give us a little time to fill everything in. Lots of work to do here, which is why we hired Andrew Mager to be a developer advocate.

1. You can think of us as S3 for location data. Before S3 scaling file storage was a messy, complex, and costly proposition. After S3 it was as simple as posting files to an endpoint. We do this for geometric/location data. We store and index your data and allow you to easily do geospatial queries against your location data.

2. We give away 1m calls a month, which is the same that Twitter gives away for free, and more than Google Maps does. The pricing plans are based on what it'd cost to manage your own infrastructure. We'll be adding lots more features, options, data, etc. in the coming weeks/months.

3. To the people saying they can just install GeoDjango and PostGIS, I have a few questions. How much time are you spending building and maintaining your own location API? How much does it cost to rent/buy that hardware? Is your location infrastructure multi-homed in three data centers? Does your location infrastructure come with premium content like global weather data and 16m+ business listings? Can it keep up with the real-time geohose from Twitter? Does it do URL callbacks and S3 backups automatically? We handle all of that for you.

4. The location marketplace will allow developers to tap into billions of points of interest and previously unattainable features for low monthly costs. For instance, Metacarta and Quova solutions are very much enterprise solutions that cost real money, but they'll be available on demand for small per-drink/per-month costs. We've indexed over half a terabyte of location data (when you're talking about point data that's a lot) that's instantly available to developers.

5. We have robust SDKs for the iPhone, including an augmented reality view. The AR view allows you to show a sophisticated AR view of your data, in our API, in about ~10 lines of code.

You all are absolutely right that we need to do better at messaging and whatnot. We're working on that. All I ask is you give us a chance and don't dismiss what we're doing as "just GeoDjango."

As someone who's built his own infrastructure, I'll respond.

1. 30m requests/month is about 11.5 requests/second. My service can do 300 search requests/second sustained off of one machine. That's with a pretty minimal setup. Data is backed up to S3. Multi-homing isn't a big priority for me, but I could easily do it for not much effort/cost.

2. If you're whitelisted, Twitter actually gives out 14.4 million calls a month for free. If not, it's 111,600 calls per month. So, I'm not sure where you're getting your numbers here.

3. I've spent a lot of time building my service and API, but AFAICT you're not providing an API I could vend directly to client applications I don't own myself. Ditto on the premium content, which I can almost guarantee you don't have the rights to give re-vending access for. (As in, I can't build a weather API service on it and charge money. Same for business info.)

4. How many of those POIs are license-free? Also, half a terabyte of point data would be a lot, if you were actually indexing that much. I'd wager $1 you're counting non-point data in that number, such as addresses and the like. That said, this is probably your strongest point,

5. That's cool. It took me awhile to write my mapping engine, and I haven't started on my AR view yet.

Honestly, your biggest problem seems to be that you don't describe your service very well. If I were you, I'd be pitching this as a good way to build up a service without the knowledge required to build a geo API in house, negotiating with geodata vendors directly (who can be extremely standoffish and unreasonable), and monitoring/scaling/etc the geodata side of things.

Also, the things I'd like to see, so that I don't have to build them myself, would be:

1. Worldwide address geocoding with no usage restrictions. AFAICT, this service does not exist.

2. Worldwide address canonicalization. If I'm missing a postal code, province, etc, or if a phone number is in a weird format, I'd love a service that would clean that up for me.

3. Address unification among services. I'd love to be able to tie together the primary ids of a given address among the various services: Yelp, foursquare, Gowalla, OpenStreetMap, etc.

> Also, the things I'd like to see, so that I don't have to build them myself, would be: 1. Worldwide address geocoding with no usage restrictions. AFAICT, this service does not exist. 2. Worldwide address canonicalization. If I'm missing a postal code, province, etc, or if a phone number is in a weird format, I'd love a service that would clean that up for me. 3. Address unification among services. I'd love to be able to tie together the primary ids of a given address among the various services: Yelp, foursquare, Gowalla, OpenStreetMap, etc.

I think #1 and #2 here are the most interesting -- with really limited providers to solve the problems. Address geolocation is a PIA, foreign address geolocation is even worse.

A non-connected private version would also be worth something (lots of money in the Government for this kind of service, usually not on the Internet).

I'm really surprised by the cost as well. I develop ArcGIS Server apps for a living, and basic software licensing to do anything interesting is equivalent to 2 months of service at SimpleGeo. And the ESRI data ecosystem is far larger than what SimpleGeo is providing.

I was hoping to see a product targeted at small businesses, but honestly I don't see the appeal. There's a great opportunity to fill in the market gap that ESRI is ignoring, but ESRI will no doubt be rolling out cloud services soon too. I guess we'll have to wait and see how this pans out.

From someone who lives and develops in the real world, your prices, current offerings, and future offerings are right on the mark.

Simply GeoDjango you are not. Nice job.

You're saying geometric data, but it's only points right? Is "GiselleDB" B-Trees on the geohash? And spatial queries are limited to 'nearby' so far, right?

I love geodjango and postgis, but it's a lot of overhead.

I've been following the simplegeo stuff, and it seems pretty cool.

Their website is missing a clear and specific description of their service.

Here's what we need front and center.

"SimpleGeo, in its most basic form, indexes data by location using latitude and longitude so that you can easily search for things that are nearby. This allows developers to easily scale, manage, and query their location data. On top of this basic feature, SimpleGeo also provides rich location data, reverse geocoding, and other tools to help developers leverage the location of their users."

http://help.simplegeo.com/faqs/company/what-would-i-use-simp...

That's still a terrible description.

What makes location data "rich"? And how do developers use "rich location data" to "leverage the location of their users?"

What exactly does it do?

I did a talk with Matt Galligan when he was in Vancouver last month. You can see the video here: http://blip.tv/file/3340542/

I can't seem to dig up a video of Matt's talk.

Basically, Navteq and Tele Atlas own the base map data. Google, Bing, and MapQuest add intelligence on top of this data with features like Street View and driving directions. Foursquare and Gowalla add another layer of intelligence with where people and places are at right now.

Rich location data is the "what,when,why,who" is at this location.

SimpleGeo is: 1) Storage for location with metadata about what's at that location 2) A marketplace to buy and sell data about locations

So, off the top of my head. I can create a door-to-door canvassing app and ask residents if they are Democrats or Republicans. Then I could sell this data to others that would be interested in this data.

Even after reading that, I still don't understand what I'd do with it. I'm guessing I'm the target audience since I'm developing a location-oriented application right now and I've developed several before, so if I don't understand what's going on, there's probably a problem.

They have a better one on their blog:

"SimpleGeo is a ready-to-use location infrastructure that makes it dead simple to make your app location aware."

http://blog.simplegeo.com/

That was my thought exactly. They've got 13 people doing what exactly? I tried to read their story, but the "Our Start" page in the company section isn't there.

It's a really clean design, but really sparse on real information.

I agree. After reading around a little I SORT of get what it does, but not really.

Wow, that's expensive: http://simplegeo.com/plans/

I just got a beta invite and was planning to spend the weekend moving my real estate listing data to the service to try it... at these prices I am happy to do it myself.

I feel like the pricing is aimed at funded startups or mature businesses, not bootstrapped startups.

The site design screams Metalab. Did they design it?

They've built a good API, but I don't know what it is either. I've been a beta user for a little while, and was originally going to use it to store bus stops in Honolulu county, to make finding nearby ones from your mobile phone quicker. In my initial testing it returned all kinds of locations near me, when all I wanted were those on my own "layer". I was getting coffee shops, and all kinds of stuff. I guess that might be the intention? It was never clear to me, and doing it locally was actually easier and faster (using geopy). Just my $0.02.

Umm... So I watched most of that video, thinking, they're going to show something cool worth $2.5k/m to a start-up soon.

Like hell they did.

$2.5k / month would buy you a lot of time with a developer doing GeoDjango work.

Not only that, but I've just realised... $2.5k/m only gets you 10 times as much as a free account. Sounds like such a great deal!

And then if you want THREE times as much as the start-up account, you have to pay FOUR times as much - $10k for 30 million calls.

Maybe they should have hired a mathematician instead of geographers....

Not all services scale perfectly linearly. They are likely limited by certain costs associated with using an Amazon infrastructure or their own infrastructure, or it is simply a matter of customer hot spotting trends and perceived demand.

Expensive! That's all I can say. I was imagining more like a $9.99/month plan.

Me, too. I've heard them compared to S3 and Heroku. Their pricing plans aren't like either of them.

I think we'll see them get changed big time, perhaps with grandfathering in of the current free plan.

they seem to be doing something interesting, but documentation and company statement is very vague. Even the demo code does not work out of the box. The video on their homepage doesn't make any sense at all..

The video was location data we aggregated during SXSW for our vicarious.ly visualization. We have lots of work to do on messaging and documentation.

What bugs did you find? We use the Python client internally and the iPhone SDK is being used in production. Maybe post on http://help.simplegeo.com?

They have niceish REST Apis, but only support OAuth for authentication. Have you ever tried out an OAuth protected API from the browser? It's bloody hard! (Google has their OAuth playground, but if there is something similar for other sites I haven't seen it)

I started using SimpleGeo when it was in Beta and I was pretty amazed to see the plans... way too expensive for a startup. I'll go back to GeoApi in a few. Sad..

This looks really cool. I'll be checking back when more of the site is up to get some more detailed information on the capabilities and the offering.

Reply on news.ycombinator.com