L'analyse de données à la portée de tous.
Une solution sur site pour les entreprises.
Une solution hébergée pour les entreprises.
We've got a fantastic guest post today from Tim Graham, a Tableau user from Australia and past winner of one of Flowing Data's visualization contests. Tim brought together several visualizations from Tableau with a diagram from another tool to create an amazing information graphic. The resulting image below groups countries of the world by their most popular brands. It's apparent that countries that are geographically close tend to share similar taste in brands, but there are some really interesting findings as well. The former USSR is more aligned with Asia than Europe in terms of what brands they prefer, while the three most "brand-aware" countries in the word seem to be the United States, Singapore, and... Trinidad and Tobago?
Be sure to click the graphic to explore it full size.
How did he do it? Tim writes:
Google Insights for Search is an application that lets you see the relative popularity of a search term across regions and countries of the world.
Figure 1 shows the results for searches on “Budweiser”.
The numbers reveal the relative geographical popularity of Google searches for Budweiser. The numbers are scaled from 0 to 100, so in the above example we see that in Paraguay, searching for “Budweiser” is about twice as popular as it is in Ireland.
Surprised? I expected the United States to have the strongest interest in Bud, and as I played with this application I found other surprising results. I started to wonder what the interests were in various countries, and which countries shared similar brand interests. The application doesn’t easily tell me the answers, but it does let me download the data… enter Tableau.
With a bit of data wrangling, I mashed up the Google Insights for Search data for each of the one hundred brands in the Business Week Best Global Brands 20082.
In Figure 2, I show the same Budweiser results using a Tableau heat map, for any countries that scored ten or more.
With a bit more data wrangling, I classified each country according to the United Nations geo-scheme so that I could compare interest in brands within and across geographical regions.
Figure 3 shows all country and brand combinations that scored highly (>68) in the Central America region.
Looking across the rows, the heat map shows that Disney is a popular search right across Central America, more so in Mexico than Panama. Colgate is a popular search in Guatemala and Panama, but not as popular in the other countries.
Looking down the columns, the heat map shows that Mexicans search for more brands than Costa Ricans. Mexico is more interested in searching for Disney, Nescafe and Marlboro than it is for Coca Cola, MTV and Caterpillar.
What does this say (if anything!) about the culture and lifestyle in various parts of the world? What does it say about the brands' marketing strategies and penetration? Why does Disney have broader "search appeal" than Colgate?
Curious, I repeated this for each region of the world, and found that brand interest varies quite a bit. There seems to be “blocs” even within the UN sub-regions. The results are shown in Figure 4. The image is big, so scroll around the world and explore.
My next question was to understand “mutual popularity of brand searches”. That is, where one brand was popular in two separate countries, I made a connection between them. I also tagged that connection based on the continent of the country that owns the brand.
After repeating for all combinations (literally thousands), I turned to a specialty network diagramming tool called yEd Graph Editor4. This allowed me to integrate my Tableau findings into a network diagram, trying to keep connected countries as close together as possible. (With thousands of combinations, the map gets real messy real fast, so I limited the results to scores greater than 68. This seemed to produce something usable. In other words, only very popular searches are included.)
The results are shown in Figure 5.
The red circles are mainly amongst red lines, green circles are amongst green lines and yellow circles are amongst yellow lines. This means that even in this era of globalisation, countries tend to search Google for brands that come from their own continent.
We also see that the clusters of circles don’t follow the UN sub-regions exactly. For example, Greece is removed from the rest of Southern Europe. Why? When we look at the heat maps for Greece and Southern Europe, we see that Spain, Italy and Portugal are more like the other “western” EU countries, sharing an interest in searching for Ikea, Duracell and Audi brands. Greece however is more interested in searching for Rolex and Sony, so it moves closer to other countries that share that interest. Other interesting examples are Bolivia and Ghana.
The heat maps are my attempt to name the clusters that showed up in the map. I think it is interesting that “culturally similar” geographic regions can be identified without any use of latitude or longitude, just relative search interest in only one hundred brands.
According to Business Week’s Top 100 brands, the world has some major “brand search hemispheres”, which correspond roughly to “Computers and Consumer Electronics”, “Aspirational Brands”, “Health and Beauty” and “Entertainment”. I’ll post an additional map showing these hemispheres at my blog - http://data.timgraham.net
There is also a general absence of asian countries such as China, Japan and Korea. I guess they either don’t search for the biggest global brands, or they search using non-latin based characters.