On Apr. 14 2014, there was the most dynamic and important match in 13-14 English Premier League. Liverpool was only five wins away from their first league title since 1990. Manchester city was 3rd on table, but there were only 4 points gap between two teams. Here's more description of this game by BBC.

As a football fan, I watched this game live on TV, and I found that shouts of spectators were fluctuating along the game's aspect. I found the match video later to capture the loudness of the sound and compared it to the game's important moments.

By using Audacity, I captured loudness of the sound 5 times a second, total 30,000 number of loudness data.
The whole dataset was visualized in R. I highlighted the bar with threshold of -25, -26, and -30. Here are 3 bar graphs that are differently highlighted.

After combining three graphs, I tried to match the data and loud moment of the game. Here's final result of it. At high-loudness moments, I could find important match data such as shot, scoring, cornor kick, free kick, and offside. I also could see how the match was getting hot. Shouts were very loud when Liverpool scored first two goals. Shouts lasted for a long time after each score because it was their home stadium. When Man City scored second goal, shout was loud but it didn't last long - maybe that was a huge sigh. You can read it by seeing sparse green bars right after City's second goal.

But the game was heated up again as Liverpool counterattacked. After Liverpool scored the winning goal, shouts lasted until the end of the game.

There are many services that provide highlight video of the game. Using this method, it can automatically recognize important point of the game and generate highlight video. There are robot journalism algorithm based on match data, but this method can include crowd's emotion into making highlight films.