Directions Magazine (DM): Let's start with the name of the company: MapLarge. To what does the "Large" refer?
Lynwood Bishop (LB): We’ve built an entirely new platform to make it easy to map and dynamically analyze very large datasets and also to deliver them to large audiences. From the ground up, we built an entirely new mapping database, tile engine, geocoder and server platform. We focused on performance with every line of code, with the goal of enabling unprecedented dynamic regional views of map data.
What does this mean? You can take a comma separated value (CSV) file of millions of customer addresses, rapidly perform a geocode, and instantly view them on top of a Google Map. Furthermore, you can also project those points on the fly into arbitrary shape layers and see calculated totals with dynamic user filters enabled. For example, LobbyData.com allows users to dynamically filter and visualize millions of political contribution records as points, or by Census block groups, ZIP Codes or counties.
Using standard mapping application programming interfaces (APIs) most programmers find it difficult to map more than a few hundred points on the client side. With MapLarge, our servers do all the heavy lifting so that users can dynamically map millions of points or shapes and view them in any desktop or mobile browser. Unlike most high traffic mapping servers, we don’t have to pre-render tiles and our servers can render each query in real-time. Users can dynamically query the database, and the underlying data can change without having to re-render billions of tiles.
DM: Your API supports EsriMaps, Bing Maps, Google Maps and others for basemap tiles and can render interactive maps via a client like OpenLayers. What's the most common combination (of basemap/client pair) in use now? How do think that combination is preferred? Do you expect a change in the future?
LB: Many of our clients use Google Maps either because it’s free or because they have a pre-existing license. Google is adding free usage limits so we expect to see more clients price shopping and choosing other providers. For high traffic applications we recommend a base map provider based on the latency of tile requests so we can deliver an experience that complements the MapLarge API speed.
MapLarge provides its own client wrapper to abstract Google and other basemap providers’ APIs. We also tie into the OpenLayers API for basemaps where we don’t currently have a direct API connection. The benefit of the MapLarge API client is that it greatly reduces the number of steps required to display data on the map with your own custom design. Using one of our templates you can map millions of points or shapes with just a few lines of code and switch between map APIs with one line of code.
DM: Do you host the user's data to be placed on the map, as well as the MapLarge Server (which handles geocoding, queries, rendering, etc.)? The client embeds the code to place the map on a webpage. And the map tiles are drawn from its provider (Esri, OpenStreetMap, etc.). So, where does the speed come from?
The speed comes from our custom database and map rendering engine. A few years ago we needed a map for a business application and the current technology couldn’t meet our needs without slow and inflexible batch build processes. So, we committed a significant amount of time and resources to develop an entirely new mapping platform and analytical database system. The result is what you see today.
DM: How do you charge for this cloud service and API?
LB: Our pricing is based on data size, traffic and users. As it turns out, the cost of mapping is not insignificant. Even Google is starting to charge for use of its basemap tiles on moderate-sized deployments. Our goal is to help organizations realize an effective return on investment whether it be a public relations event to increase brand awareness or increased efficiencies in a business intelligence (BI) application.
Currently, our sales team is focused on larger customers with big curated datasets. However, we are planning to release a self-service mapping portal that will be more economical for smaller users. The mapping portal will have a simple upload process similar to our online geocoder.
Persons Living with an HIV Diagnosis / 2008. Rate by State, Overall. Click for larger view. Credit: AIDSVu ( and the Rollins School of Public Health at Emory University
DM: Your latest project, AIDSVu, hosts county level AIDS data for the entire U.S. with tools to slice and dice it in many ways. What were the particular challenges of building this app?
LB: With AIDSVu, you have an organization whose stakeholders have been working for years to collect these data to build this important public health message platform, and they wanted to be very precise in the message they delivered. The MapLarge API allowed them to totally customize the look and feel of each element of the map. We spent a lot of time making sure the presentation was exactly the way they wanted it. Since we own the entire technology stack, we can make modifications to the API to serve a particular client need.
In this implementation we made several small enhancements in our clustering and shape labeling algorithms to make sure their map “felt right.” The translation between conceptual art and dynamic application was one of the most fun and challenging aspects of this project.
DM: What enhancements do you expect to add in the coming months?
We are also doing some exciting work with real-time demographics and we should be announcing some big location analytics products and partnerships in the coming weeks.