May 11, 2018

The 6 Steps of Social Media Data Analysis for Retail Real Estate

How can social media data provide valuable location intelligence to retailers and restaurants? What does the process looks like?

In this post, we'll walk you through the 6 basic steps of our social data analysis.

social data around the world

1. Gather the data. Individuals post about their interests, personalities, and behaviors on a variety of social platforms every single day. This occurs across communities, neighborhoods, and cities, creating a massive cloud of data that is loaded with authentic human behavioral information. Plus, with geotagged social media, all of this behavioral data can be tied to specific locations. We gather, filter, and clean all of this data to prepare it for analysis.

machine learning process

2. Categorize and Quantify. Using artificial intelligence, we are able to categorize and quantify this data into 100+ social segments - tied all the way down to the block level. In other words, we identify patterns and group data points that share similar traits, but we do this with a complex machine learning modeling approach.

 

categorizing segments

3. Identify and Label. After we have created these AI-generated segments, we need to interpret and identify what they mean. This is where ethnographic principles come into play. We study the human behaviors exhibited from each segment and distill their essence. In this case, the segment A_00 had a high frequency of health-related topics, which could be generally labelled as "Wellness."

 

location analysis

4. Identify the most significant segments (and their impact) for a given location. This is where the data is applied to optimize real estate success. For any store location, we compare these hundreds of segments to the real business performance, revealing whether they have a positive or negative impact. Understanding the impact of each relevant social segment on your store location allows you to optimize marketing campaigns to precisely target the right consumers or avoid prospect locations where negative segments are prominent.

Note: We use predictive modeling to identify causation, not just correlation, between segments and business revenue. 

 

custom social indicator

5. Calculate the CSI for any location. We combine all of these segments into a singular Custom Social Indicator (CSI), which reflects how well a location is likely to perform based off the social data around it. We do this across many different locations for easy performance comparison.

 

 reports and visualizations

6. Generate Custom Reports and Visualizations. Finally, we consolidate and deliver everything in a way that makes it simple for companies to actually use the data. Some of the ways we do that are with:

  • Heat Maps. Ideal for sharing the data during internal presentations.
  • Custom Reports. See all of the results at a glance. Plus, we have designed them to work alongside any other data analysis. 
  • Insights Packets. Stories and insights to inform customer experience and marketing decisions.

 

Want to learn more? See how we used this same process to predict Payless store closings with 82% accuracy. Or contact us at retail.spatial.ai.