Built a k-nearest-neighbor classifier that guesses whether a song is hip-hop or country, using only the numbers of times words appear in the song’s lyrics. Our dataset was a table of songs, each with a name, an artist, and a genre. To predict a song’s genre, we had some attributes: the lyrics of the song, in a certain format. We had a list of approximately 5,000 words that might occur in a song. For each song, our dataset told us how frequently each of these words occur in that song. We classified a song by finding the ‘k’ songs in the training set that are most similar according to the features we selected. We called those songs with similar features the “neighbors”. The k-NN algorithm then assigned the new song to the most common category among its k neighbors. To implement this algorithm, we defined similarity between songs as the Euclidian distance between them when we plotted their features in a scatter diagram. For n different features, we simply computed the difference between corresponding feature values for two songs, squared each of the n differences, summed up the resulting numbers, and took the square root of the sum. The test-set accuracy of our classifier was 74.16%, with limited feature engineering.