How we use Open Street Maps for more effective geo-contextual targeting at MiQ

Nitin Vinayak Agrawal, Data scientist II, MiQ & Tamphaleima Laishram, Product analyst I, MiQ

As programmatic advertising evolves towards a cookieless future, it’s important to stay ahead of the curve and adapt to privacy-centric forms of targeting and optimization. Marketers know that first-party CRM data can be foundational to an authenticated ID strategy — but it won’t give you 100% of your audience strategy. Brands still need to scale audiences in order to drive the best return-on-ad-spend (ROAS). A great way to do that is with geo-contextual datasets.

What is geo-location data?

As our previous blog post on geo-contextual targeting explained, geo-location data is an aggregation of datasets tied to locations, such as postal/zip codes, which provides insight into the people who live in a particular place. It works on the idea that people who live in close proximity to each other are more likely to share similar attributes. Just think about neighborhoods in your city — can you identify areas with similar household incomes, interests, jobs, life-stages, etc?

When it comes to programmatic targeting and optimization, we can use it as a cross-channel identity signal that persists across DSPs in the granular impression data that you get at the log-level.

Sounds like it could be a bit too all-knowing — is it?

The opposite is true. Rather than the one-to-one hyper personalization of some forms of consumer data, geo-location data is always viewed in aggregate. That means it’s anonymous data and an individual person’s data is never gathered in any form. This is why it is central when thinking about cookieless targeting.

Geo-contextual datasets at MiQ

As a global business, MiQ has a variety of regional, geo-contextual data sets that we can tap into, but one unique dataset that we use is called Open Street Maps (OSM). This dataset provides detailed information on the characteristics (e.g. tourism, historic, landuse) of entities (e.g. shops, buildings, offices)at any given location. MiQ decodes this data to help advertisers select the area with the highest potential for successful engagement with their target audience.

Read on to learn how we use OSM.

Four basic elements of OSM

OSM consists of four basic elements: Nodes, Ways, Relations, and Tags.

Nodes are geographic coordinate points defined by latitude and longitude coordinates and id numbers e.g. MiQ’s Bangalore office is a node with lat long coordinates and different tags.

Ways are ordered lists of nodes that define a polyline or linear feature, such as a road or area boundary e.g. our Bangalore office building Skav 909 is made up of multiple nodes like MiQ, Kaze Bar, Lexus showroom etc.

Relations are ordered lists of nodes, ways, or other relations and are typically used to denote relationships between two or more elements e.g. Hudson City is a relation made up of several eateries, churches, parks and other nodes.

Tags are attributes of locations of Points of Interest (POI) e.g. landuse, amenity and name. Major POIs include shops, offices, leisure sites, highways and boundaries.

Exploring Hudson City through Open Street Maps

Using OSM, we gain valuable insights into specific locations and their defining characteristics. Take, for example, Hudson City, a popular residential area in New York with top facilities and therefore a high cost of living. According to data, this area has a high affinity towards luxury car brands such as BMW.

Using OSM data to analyze amenities

Looking at OSM data for the zip code 12534 in North America, we can gain insight into the various amenities present on Columbia Street in New York. OSM provides detailed information on eateries, parks, places of worship, restaurants, and more, allowing us to better understand the local area and target accordingly. Some of the key features we can identify are listed below:

​​How to access OSM data

Open Street Maps (OSM) is a crowd-sourced and open-source data set that can be accessed in several ways. The most common methods are using the Overpass API or direct bulk downloads from GeoFabrik.de or BBBike.org in raw format (.osm).

Why we choose OSM over Google Maps

Google Maps may be a more well-known map provider, but OSM offers distinct advantages. OSM is crowd-sourced and freely available. Google Maps requires API access from Google Cloud Platform (GCP), charges above a mapload limit and Google captures geographic information using its private technologies and survey tools. Other differences between the two can be found in this Medium blog.

Geo-contextual targeting at MiQ

In the post-cookie advertising world, geo-contextual datasets are crucial for understanding audiences’ behavior through their geo-interactions. OSM provides an additional set of features that allow us to infer audience interests based on geographic location. This helps in planning geo-based strategies by providing urban/rural information and other location characteristics. At MiQ, we’re always looking for new ways drive performance for our clients whilst minimizing cookie usage — a focus on perfecting our geo-contextual targeting is just one of these.

For example, luxury brands can find potential user bases by targeting locations in and around tech parks, which are typically urban areas that host higher-income earners. With MiQ’s team of 300+ data scientists and analysts, we have leveraged this kind of map data to:

• support anonymous precision targeting
• scale audience reach
• showcase relevant audience interests
• group and classify audiences based on their location behavior

Considerations for using OSM data

• It’s important to keep in mind that since OSM is crowdsourced data, generalizing for the whole country may not be accurate, especially for rural areas
• Additionally, studies have shown that OSM is approximately 83% complete, so it may not be the only source of location data that advertisers should use

However, in contrast to other mapping tools, OSM is freely available and crowdsourced, making it more accessible. OSM also provides additional coverage with infrastructural features on top of census and third-party data sources. For these reasons, we continue to use OSM in our geo-contextual strategy, combining it with other data sources for a complete view.

This strategy supports the cookieless toolkit that we have developed at MiQ, which encompasses a variety of solutions including our data-backed Identity Spine, to help brands receive future-proofed audience insights and improve their targeting and performance.

Interested in using geo-contextual targeting to drive better cookieless performance? Contact MiQ today to learn how we can enhance your audience reach and engagement using OSM and our other data sources.

Nitin is a data scientist working in MiQ’s Bangalore office who spends time outside of work watching web series and playing poker. Working in the same office, Tampha is a product analyst at MiQ who enjoys watching historical documentaries and is a Tolkein fan!