Data at Work

The location data industry is using multiple sources of data with a varying degree of quality. This means suppliers often have a significant amount of overlapping records. While this does not impact quality or validity of the data, it poses a challenge for buyers who need unique values for their analyses.

Most off-the-shelf vendors leave this unaddressed as it makes their data volumes appear larger. However, when buyers prepare the data for analysis by removing overlaps, they often see a significant drop in overall volumes. 

Image attribution: Storyset via Freepik

Earlier this year, we launched the public beta test of our proprietary POI collection and verification app, Geolancer. Aimed at providing up-to-date and manually-verified POI datasets, Geolancer is already helping ride sharing and real estate companies enrich their platforms with a hyper-accurate POI data feed. In this article, we provide a peek behind the curtain and explain how we maintain the high quality of POI data in Geolancer.

The Covid-19 pandemic and following global lockdowns were devastating for the retail industry. Activities like shopping and leisure quickly fell as people’s focus shifted to dealing with a pandemic. Essential retail like grocery, drugstores, and homeware had to adapt quickly to accommodate the needs of the hour, while non-essential retail segments of fashion, sports, beauty, luxury, etc. went into complete disarray.

Location data is collected from multiple sources of varying quality GPS signals from mobile devices, beacons, and WIFI connections, the notorious Bidstream, and more. In most cases, even genuine location data cannot represent the entire population of the region. This discrepancy can be attributed to smartphone penetration in the country, app-specific demographic variations, hardware inconsistencies, and sources of location data.

To perform meaningful analysis that accounts for mobility patterns and other trends in a larger region, data scientists use projection models to make an accurate estimation of a region’s population and normalise data counts to fit the use case. This is called data extrapolation.

“It is an undeniable statement that programmatic advertising is going to take over.”

That is what Quadrant Data Scientist Roger Ganga told an audience at company headquarters in Singapore during our latest network event. And the facts speak loudly for themselves: the proportion of marketing campaigns run through programmatic advertising soared from 17 percent to 28 percent in just a year.