Anomaly/Outlier/Novelty detector -- 2

Hi there! We are looking for a data scientist freelancer to develop an outlier/anomaly/novelty detection model. We'll provide a significant dataset of not labeled data hence the use of unsupervised techniques is a must. The acceptance criteria will be the following: - As it has been metioned, unsupervised methods are a must for the detection - We expect that the outliers founded by the model could be reasonably explained - The model will be tested with data that it has not seen before Deliverables: - A sheet with the detected outliers by the model for the provided data. This sheet should have the columns: id_row, number_column, why (e.g. Column 6 is drastically far from the mean of the data) - Notebook or python script with all the steps in order to get the results - Conda enviroment file .yml used test data can be readed with pandas.read_pickle("[login to view URL]",compression="gzip"). The returned object is a pandas.DataFrame.