Our Best City Contest shortlist is out:
Vote for the $10,000 winner!

Best City: Hong Kong

Approach: To complement the EIU’s existing liveability index with an awareness of a new indicator, a city’s spatial characteristics. This new category seeks to account for spatial aspects of city life: urban form (sprawl, green space),geographical situation (natural assets, isolation, and connectivity), cultural assets and pollution.

Best City: Calgary

Approach: Two new categories, Environment and Life Satisfaction, were defined and indicators calculated with variously sourced data. The visualization is an interactive interface and can be found at www.torre.nl/bestcity. For each city, a “radar” chart is visible which shows the different categories of the index, the bigger the chart, the better the city scores.

Best City: Atlanta

Approach: According to the EIU’s linearly scaled method, two hypothetical cities that score of 2 and 2 (out of 5), and 3 and 1, respectively, would get the same average. However, one would likely rather stay in City 1 than City 2, since in City 2, a score of 1 is “acceptable” and 3 is “uncomfortable,” while in the former, both scores of 2 are “tolerable.” This entry scales the subcategories quadratically, progressively penalizing as scores move from 1 to 5, thus ensuring that City 1 would rank higher than City 2.

Best City: Zurich

Approach: The EIU assigned coefficients to particular columns and calculated weighted averages. This method heavily depends on point of view and is inherently arbitrary. This entry does not introduce any weights or arbitrary coefficients, but calculates ranking from the database itself. Technically speaking, it reduces the complexity of data: combining as many parameters as possible to a single score (also known as Data Dimensionality Reduction).

Best City: Geneva

Approach: This model factors in the fact that people from cities with different liveability condition values various attributes differently. Unlike most models that look at livability from the perspective of the city and what the city can offer, this model evaluates livability from the perspective of an average city dweller, their expectations and their experience through an average day. This model assumes the best cities are not cities that create maximum satisfaction but cities that cause minimum dissatisfaction.

Best City: Determined by user input.

Approach: This entry is based on 1) EIU data, 2) Hofstede cultural analysis and 3) official language(s) per city to help anyone to find out which is the best city for them. For example, the contest entrant’s best city in the word are Toronto and Montreal, since he speaks French and English, his cultural background is French, but he is more risk-prone and willing to work in organizations with flatter hierarchical relationships.

Best City: Determined by user input.

Approach: This entry dynamically computes the best city for a given user’s set of preferences. The preferences from a set of users can then be aggregated to produce a ranking that best approximates the collective preferences of everyone in the group. The site completes the loop by automatically updating a public data set of preferences on BuzzData. This offers open access to a continually growing storehouse of information — individual preferences about factors that make a great city from users around the world.

Best City: Vienna

Approach: This index takes all the EIU’s variables then uses a k-means clustering algorithm to divide these into clusters of similar cities. The k-means process is repeated many times with different centres to give a measure of how similar each city is to each other city. This entry makes it really easy to change the sets of “good” and “bad” cities – say you have a set of cities that you find are good for short-term travel and others that are terrible for this, it is easy to take these sets of good/bad then use them to create a new index.

Best City: Determined by user input.

Approach: Rather than try to determine a single score or combination of scores, this entry lets the user choose and explore the entire data set themselves. Using maps integrated with the EIU data (as well as MotionCharts with the each of the 140 cities and all of the indicators plotted, the user can select and compare cities one-on-one, a selection of a few cities or just explore the entire dataset for themselves.

Best City: Determined by user input.

Approach: The single most important factor in determining the best city in the world for you is your preferences, which are unique to you. This visual city selector pinpoints your personal best city by letting you choose an area of the world, then specify your priorities using the factor slider to adjust their relative weights.