Crime Prediction in San Fransisco

In June 2015, Kaggle began a competition named “San Francisco Crime Classification”[8], ending in June 2016. The competition’s dataset caught our attention due the subject being very tangible, with crime being at the forefront of modern media and to San Francisco being culturally significant due to its current tech industry. The dataset is also described by geographic and temporal features, therefore enabling potentially interesting visualizations. After our initial investigation of this dataset and the Kaggle competition, we realized that there was a large amount of accessible information on different ways of analyzing this very data set through blog posts and scripts published for this competition on Kaggle and Kaggle’s forums. Through the nature of Kaggle competitions, a means of evaluation is also provided for by the competition rules. We chose this dataset for these reasons. The “San Francisco Crime Classification” competition and its accompanying dataset, provided by SF OpenData, consists of 878,049 samples of crime reports from all neighborhoods of San Francisco spanning from January 2003 to May 2015. The data is initially split by Kaggle into two sets: the training and testing set. Odd numbered weeks (1, 3, 5, 7, . . . ) are put in the training set, and even numbered weeks (2, 4, 6, 8, . . . ) are put in the test set. The fields in each sample point are given in table 1.