lost my dog said:
datageek said:
April 16th, 2019 Krispy Kreme opened with lots of visitors. Many wondered if the four stores of Shipley's would see any changes. So far only the Southwest Parkway showed a temporary drop in traffic and rebounded. It seems there is plenty of room for Krispy Kreme to join in on the donut fun.
Check out the foot traffic here: https://cl.ly/7d13fd3aa616
So I regularly get Google asking me to rate my transaction at a business which is next to the business I actually patronized. Geolocation is not perfect. How do you account for this in your analyses?
As an imperial skeptic, I was highly unsure of the accuracy of the data sources at first. I started by pulling data from public events where the headcount was already well documented (stadium events). Next, I pulled foot traffic for restaurants where they tracked all reservations and had them compare their data to what I had, and finally the same thing with restaurant/retail where the volume was low to test the lower bounds of the data.
Not all locations are we able to purchase foot traffic data. For accurate foot traffic information, data is needed and collected from WiFi, GPS, Beacons, and even user motion. Panel biases are modeled and accounted for in extrapolations. If there isn't enough 3rd party data to put it all together and make it statistically significant, we can't show anything meaningful.