Overview

A Chinese technology company specializing in smart mobility and automotive intelligence wanted to help its client. They traditionally provide industries with HD maps, navigation, location-based services, and autonomous driving solutions. Their client was a retail tea manufacturer seeking to enter the Asian, North American, and European markets, starting with Hong Kong as their first retail expansion target.

The company partnered with Xtract.io to obtain a reliable POI dataset for Hong Kong, with a focus on cafés, restaurants, malls, and retail outlets. Their goal was to understand the retail landscape in detail, mitigate risks associated with brand expansion, and guide their client in making a successful entry into the Hong Kong market.

Challenges

We were assigned a 50×50-mile section of Hong Kong’s map and had to identify every relevant Point of Interest (POI) within it.

• Given the client’s large area of interest, it was necessary to develop a methodology to collect an extensive database of POIs, along with detailed attributes for each POI.
• Information about businesses was fragmented among several sources, including online directories, review sites, and outdated local listings.
• Many of these listings were incomplete and inconsistent, missing ratings, reviews, or important merchant details.
• The retail dataset needed to be relevant and accurate, as it is intended to inform strategic retail expansion decisions.

Alongside these challenges, we also had to maintain the quality of the POIs that we collected while ensuring consistency across hundreds of sites.

Solution

We divided the Hong Kong map into 20 × 20-kilometer polygon grids to cover the city in an organized way, ensuring completeness.

Step 1: In each polygon grid,  points of interest aligned with the client’s objectives, such as cafés, restaurants, malls, and retail stores, were concentrated.

Step 2: We identified duplicates and consolidated them into reliable data through an automated matching process.

Step 3: We collected POI details from a variety of sources, including directories, review platforms, and local listings, through a standardized procedure.

Step 4: The data was validated, and the POIs were enriched with attributes, including merchant name, address, geocoordinates, category, business type, parent merchant, ratings, and reviews.

Step 5: As a final step, our GIS experts standardized the retail datasets manually to correct errors, fill in missing details, and verify the information against trusted sources. This ensured that the collected data was accurate, comprehensive, and ready for use.

Results

Each POI was enriched with the required location metadata for targeted retail expansion, providing the client with a comprehensive view of the retail landscape. Only reliable data that fell within the specified boundary and aligned precisely with the requested lines of business (LOBs) were retained. We provided:

• Accurate merchant identification with standardized names and parent–child relationships.
• Verified location details such as addresses and geocoordinates for precise mapping.
• Grouped outlets into clear categories, such as cafés, restaurants, malls, and retail stores.
• Included POI attributes such as ratings and reviews to better understand customer preferences and gain insights into sentiment.

We delivered a rich dataset of 14,868 Points of Interest, including 1,400+ important retail and food locations. This helped the client to identify high-potential areas for retail expansion, refine their strategy, and target investments toward the most promising growth opportunities.