time-series-clustering

K-Nearest Neighbor, and hierarchical clustering approach to Time Series Clustering on Stock data:

• The data is an unlabeled collection of daily variations for a singular stock logged every minute
• There are 277 days worth of data and 1440 data points, corresponding to one observation every minute
• All days have an equal number of data (1440) and no data point is absent

Here is the given data after cleaning

Generated clusters

Do checkout the indepth python notebook for the steps and logic behind those steps taken to perform this task

KMeans Clusters

KShape Clusters

Complete Clusters

Ward Clusters