7-Eleven vs. Circle K: Where does one have more stores than the other
Visualising spatial distribution of the two largest convenience store chains in Hong Kong
TL;DR
7-Eleven and Circle K are two major convenience store chains in Hong Kong. Where could you find more store in one chain compared to the other? And what is the magnitude of differences?
Note: Chinese version of this article is also available in my Medium (i.e. here!).
Hong Kong is packed with two major convenience store chains, 7-Eleven (7–11) and Circle K (OK). While there are nearly 1,000 7–11 stores, there are only 350 Circle K stores in the whole territory.
This brings us two questions:
- How dense and concentrated are these two chains?
- How dense is 7–11, as compared to Circle K?
To answer these two questions, I scrapped the location of convenience stores from the official webpage of the two chains¹ to map and visualise their density.
Disclaimer: The dataset is not fully validated and may not be updated to the latest version.
Mapping where there are more 7-Eleven/ Circle K
Here’s an overall distribution of 7-Eleven and Circle K stores. The map shows the differences between the number of stores of the two chains, binned up into 0.125 sq.km. (12.5 ha) hexagonal grids/zones. The hexagons have a diagonal of around 440 m².
These hexagonal zones are first coloured by whether there are more 7-Eleven or Circle K inside them. The colour refers to the one in the logo of these two chains. Green hexagons imply more 7-Eleven are located in the zone, while red hexagons indicate more Circle K. What if both chains have the same number of stores? Fill the hexagon with grey! Finally, how about the area without hexagons? It means there are neither 7-Eleven nor Circle K in the zone.
The hexagons are then scaled by the differences between the number of stores. A large green hexagon means there are much more 7–11 than Circle K in the grid, while a small red hexagon means there are one or two more Circle Ks in the zone.
What does the map imply about the distribution of these two chain stores? Let’s zoom in to a few areas.
Central and Tsim Sha Tsui
7–11s are very dense in these two areas — much denser than Circle K. They are reflected by the large green hexagons sticking together.
Near Lan Kwai Fong, there are only about 4–5 Circle K stores. However, there are nearly 30 7–11 stores. When you want to grab some late-night munchies from 7–11 yet drunk enough to miss the store at the corner of the street, you still do not take an u-turn — take a few steps forward and you will find another 7–11 in front of your eyes, situated in the middle of the street.
Nathan Road
Among Nathan Road, whether you are in Tsim Sha Tsui, Mong Kok, or Shum Shui Po, you could find yourself surrounded by 7–11.
Residential Area
The red hexagons are mostly smaller than that of the greens. Put it in another way, even zones with more Circle K than 7–11 exist, the differences are not that huge. There are mostly 1 or 2 more Circle K in those zones. While we could see there even have nearly 10 more 7–11 than Circle K in business districts, the differences are much, much smaller.
Usually, you could find more Circle K are in residential areas, especially those further away from the downtown. In the zoom-in map above, the red hexagons are located near the hill of Tai Wai, which is far away from the Sha Tin city centre. Meanwhile, in Kowloon, the red hexagons are in Tsz Wan Shan and Sun Po Kong.
Interestingly, the largest red hexagon (i.e. the zone where the Circle K is the most dominant store) is in Wan Chai. There are 8 Circle K stores and 5 7-Eleven stores in that grid.
The proportion of two chains, by district
What are the differences between the proportion of 7-Eleven / Circle K stores among 18 districts?
The bar chart above shows the proportion of the two chains of 18 districts, ordered by the percentage of 7–11 stores, from largest to smallest. The number in the middle of each bar represents the number of 7-Eleven / Circle K stores in each district. Say, the 76 and 11 in the first row implies there are 76 7-Eleven and 11 Circle K stores in Central and Western District. The proportion of the two stores is around 8:2.
The overall 7-Eleven / Circle K store ratio of Hong Kong is about 73:27. When grouped by districts, the ratio ranges between 8:2 to 6:4. Circle K never takes the majority. When there are nearly a thousand 7-Eleven and merely around 350 Circle K, this is never a surprising result.
As we have seen in the map above, the expansion of 7-Eleven in commercial areas is much more aggressive than Circle K. The three districts with the greatest proportion of 7-Eleven are Central and Western (81.7%), Shum Shui Po (81.3%) and Yau Tsim Mong (78.9%). Meanwhile, districts with a smaller proportion of 7-Eleven stores are districts with mostly residential areas. The three districts with the smallest proportion of 7-Eleven stores are Tuen Mun (63.5%), Wong Tai Sin (63.6%) and Southern (63.9%). This trend is also reflected on the map — there are more red hexagons in Tuen Mun area.
What are the things missed?
The bar chart and map above are merely some fundamental spatial data wrangling and visualisation. Indeed, there could be tons of follow-up base on the map.
There’s an observation (hypothesis, to be exact) from my friend that it is easier to find a Circle K store in public housing neighbourhood. Eyeballing the map above, it seems there are more red hexagons in areas further away from the town centre and near the hills. But are the differences statistically significant? That will be a research question for those living in Ivory tower.
One additional note: the business model of 7-Eleven and Circle K are quite different. Circle K stores are usually larger with various snacks available in one single store. The display cooler storing the cold beverages have 3–4 columns. Meanwhile, the size of 7-Eleven are usually smaller — the corridors are narrower, and the coolers are quite tiny. If we use gross floor area (GFA) to summarise the total of each grid (instead of the number of stores), the results could be different (if it is possible to get the floor area of each convenience store, of course).
Finally, VanGO is also one of the convenience store chains. I excluded it as there are less than 50 VanGO in Hong Kong now. Besides, visualising three variables within a single map is quite challenging³. So, let’s keep the story simple and assume VanGO does not exist in this analysis, shall we?
[¹]: As a useR, I use rvest to scrap and transform the webpage text to spatial data.
[²]: That’s exactly the grid I used in my old article Gridify Hong Kong.
[³]: If you are mad enough, try to make a trivariate map to visualise the proportion of each chain. John Nelson has a crazy tutorial for you ( https://www.esri.com/arcgis-blog/products/arcgis-pro/mapping/crazy-expression-symbology-and-crazier-cmyk-maps/). I am more than welcome to share the convenience store data if you are going completely nuts.
