Introduction

The history of board games can be traced back to 3000 BC. Unlike video games, it doesn’t require so much technical support. Many board games are simulations of reality. Nowadays, although video games are experiencing the best time with the support of fast-developing technologies, there are still so many good board games coming out every year. Board Game Geek is a professional board game website. Users can find every single board game and its related information on it. The information includes the game’s introduction, reviews, user ratings, professional ratings, price, where to buy and etc.. According to Board Game Geek, Board games can be categorized into 84 types and 54 mechanics. I personally like to play board games.

Through this project, I want to discover the history of board games and find out what the most popular board games are from different perspectives. The discovery can be proceeded by answering specific questions about the dataset. For example, what category has the most games? How many games of a certain category came out for a certain year? Is there any correlation between the game mechanic and the game time? Are the top ranking games sold well?

The project output is a data vis dashboard composed of 4 graphs regarding 4 questions asked above.

Examples & Inspirations

I looked up for some data visualization examples that match the board game topic. Since the topic is very particular, it’s hard to find some really good examples. However, I found some good visualization design practices I really like from Pinterest.

This example demonstrates the feeling of the author regarding different types of music. I really like how the author uses the color and the shades. The dark background and bright colors draw the audience’s attention immediately. The circles make it look vivid and clickable even though it’s not interactive. Although the mood cannot be measured with numbers, it expresses pretty well by building contrast for all the elements.

This is a data vis of global carbon footprint. The legibility of the details is great. Even on canvas with this size, the details are still readable. All the countries form a foot to match the title. I think the circle is a really versatile shape that can fit in a lot of situations. The problem of the poster is that so much text has overwhelmed the limited space. There is a mess in the middle of the canvas where it is hard to tell which country is which circle. But this problem can be easily fixed on an interactive visualization.

This is another beautiful data vis poster. It has a 3D look which actually doesn’t have any meaning. But it does make the simple bar chart more impressive. I think it’s good to add some 3D effects to my future projects.

I got some design inspirations from examples. The good design can endow the graphics with life. More specifically, in the digital era where people are used to well-established web elements such as CTAs, the smart use of analogy will be great for interactive data vis projects.

Dataset, Tools & Process

For the dataset, I tried to grab data from Board Game Geek website but there is no way for an ordinary user to do it. So I googled and luckily found the open data posted by @mrpantherson from Kaggle which is a very good open data website. This board game data has 20 columns: Rank, URL, Game ID, Name, Min players, Max players, Average time, Min time, Max time, Year of release, Average rating, Geek rating(Professional rating), Number of votes, Image URL, Age, Number of BGG members owning the game, Category, Designer, Mechanic and Weight.

The data is very complete and errorless. However, one game can have multiple mechanics and categories. To make the data recognizable to the machine, I used OpenRefine to split category and mechanic into multiple columns and transpose them back into one column.

After getting the data ready, I used Tableau Public which is a very powerful data vis tool for the visualization work.

The first graph demonstrates the numbers of games in each category. Each bubble represents a board game category. The size of the bubble shows how many games are there in each category and the color is used to increase the contrast. From this chart, we can tell that card game is the biggest board game category. To make this graph, I didn’t put anything on horizontal and vertical dimensions. Instead, I put category on Color and Label, and distinct count of game titles(name) on Size. Although some of the text is hidden due to the small size, it can be seen by hovering the bubble.

The second graph shows how many games in each category were released each year before 2018 (the latest update of the data is Jun 2018). I also specify the range of the games to top 50 to see the trend. Columns represent years and rows represent categories. The size and color represent the numbers. From this chart, we can tell that all top 50 games were released in the 21st Century. Fighting, economic and adventure topics have become increasingly popular. Even though card games are the biggest throughout the board game history, they don’t make it into Top 10 in this graph.

The third graph shows how many BGG users own the top 50 games with the columns representing numbers and the rows representing game names. Colors and the order show the ratings from high to low. The interesting thing is that some really high-rated games only have small numbers of owners. The 48th game Codenames has the most owners among all top 50 games. Gloomhaven as the best ranking game doesn’t even have half owners as Codenames’.

The last graph shows the relationship between average game time, mechanics and average ratings. I put the game time(min) as the horizontal dimension and game mechanic as the vertical. Colors respresent the average rating of the mechanics. Obviously, the Hex-and Counter games take more time to play. One thing interesting is that there seems to be a hardly visible trend that the longer the game is the higher rating it gets.

Reflection

Through the project, I find it difficult to deal with different types of data. The dataset I used is not like any governmental data which has a predictable linear correlation between time and other measurements(Population, sex ratio, etc.). It isn’t like video games which the game types change through the technology innovation by the time either. It is hard to use time as a variable. Since the release year of each game is only relevant for that single game, there is no way to create the line chart based on time. I think I will discover more interesting visualizations regarding this tyoe of dataset.

There are cells that contain multiple values in the original data. I only transposed the necessary ones(category, mechanic) instead of all. So I can spend less time waiting for OpenRefine to proceed. For a big dataset that has too many distinct records, the proper use of filters can help clean up the visualization. On 3 of my graphs, I set up the filter to only show the top 50 games of all to keep the graph short.

For the design, I like to use colors. I think colors don’t always have to have meanings, you can just use it as the decoration. For my first graph, adding colors makes it look and match the topic better.

Tableau is very powerful and smart. I hope I can use it for more complex projects and ideally use codes to make advanced visualization.